|
I looked at my program for two (2) days straight. I looked at it night and day and could not make heads or tails of the problem. At one point I thought I was going to lose my mind. I am not sure I that I understand your explanation completely, but thanks. I really needed some sort of explanation so I put my mind at rest. Thanks.
|
|
|
|
|
The best part about this is that if you make it work, i.e. don't increment count before checking values again, the entire content of the if statement simplifies to
count++;
Console.WriteLine("count = {0}", count);
... because you are testing all combinations.
I'm not sure what you're trying to do but I'm sure the code is not doing it.
|
|
|
|
|
I was also reading online and saw that many people had this same problem. One of the experts said it was a dot.net error and that people should contact Microsoft tech support for a fix.
|
|
|
|
|
This is indeed a very strange problem to experience. I ran the entire English alphabet all at once and did not get an error. I ran only one special character and got errors every time.
|
|
|
|
|
That's what I thought at first, but it's not quite that simple. If condition 1 is met, you move to the next index, and test condition 2 against the new index. Therefore, it's possible to increment the indexer at least three times within the same iteration of the loop.
For example:
0 1 1 0
^
(condition 1 is met)
0 1 1 0
^
(condition 2 is not met)
(condition 3 is met)
0 1 1 0
^
(condition 4 is met)
0 1 1 0
^
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
That's true as it stands but I'm pretty sure that's not what's intended, because incrementing count multiple times can easily result in an out of range exception.
|
|
|
|
|
Hello all,
I am using sqladapter and update command in it. I am checking how update command does the functioning. For reading the command i am using getupdatecommand.commandtext and the command which comes to me is
UPDATE [empdata] SET [name] = @p1, [id] = @p2, [salary] = @p3 WHERE (([name] = @p4) AND ([id] = @p5) AND ([salary] = @p6))
where name,id and salary are the tables in db. I am not able to understand what these @p are and how is all this working. I will be very thankful for its explanation. 
modified 6-Mar-14 15:02pm.
|
|
|
|
|
The @name stuff is a parameter, and it allows you to transfer information from your C# code to SQL without converting it to string, or concatenating strings. This is important, because not using parameters can be very dangerous.
Suppose you have a string you want to update against an integer id:
using (SqlConnection con = new SqlConnection(strConnect))
{
con.Open();
using (SqlCommand com = new SqlCommand("UPDATE myTable SET myColumn1='" + myNewString + "' WHERE Id=" + myId.ToString(), con))
{
com.ExecuteNonQuery();
}
}
Will work, but it leaves you wide open to something called SQL Injection Attacks, where I can use the value in the string to add commands to the UPDATE: if the myNewString variable contains
Hello';DROP TABLE myTable;--
Then what gets passed to SQl is a comannd string:
UPDATE myTable SET myColumn='Hello';DROP TABLE myTable; Which as far as SQL is concerned is two valid SQL commands and a comment.
It updates every row in your table to "Hello", and then deletes the table.
Parametrized queries are a way to avoid that.
using (SqlConnection con = new SqlConnection(strConnect))
{
con.Open();
using (SqlCommand com = new SqlCommand("UPDATE myTable SET myColumn1=@MYSTRING WHERE Id=@ID", con))
{
com.Parameters.AddWithValue("@ID", myId);
com.Parameters.AddWithValue("@MYSTRING", myNewString);
com.ExecuteNonQuery();
}
}
Passes the same info to SQL, but the string value is now passed as a parameter and not interpreted as an SQL command at all.
It is a very, very good idea to use parameters at all times - it prevents other problems as well!
[edit]Typos - loads of typos... [/edit]
Those who fail to learn history are doomed to repeat it. --- George Santayana (December 16, 1863 – September 26, 1952)
Those who fail to clear history are doomed to explain it. --- OriginalGriff (February 24, 1959 – ∞)
|
|
|
|
|
Have you just given your email address to a Danish member[^] who joined 11 years ago and has never participated since? 
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
Sh*t d*mn b*gg*r f*ck.
Those who fail to learn history are doomed to repeat it. --- George Santayana (December 16, 1863 – September 26, 1952)
Those who fail to clear history are doomed to explain it. --- OriginalGriff (February 24, 1959 – ∞)
|
|
|
|
|
Very very thankful for your so clear explanation Also please explain what here AND operations are doing, i will be very thankful to you
|
|
|
|
|
AND is what is called a boolean operator, along with OR and it is used to combine two "tests" to make a bigger, more powerful test.
AND is true is both the test on the left and the text on the right are both true: so
Object is round AND object is blue
is true only for blue balls - it is false for red balls, blue cubes, green pyramids, ...
OR if true if one or the other of the left and right tests are true, and false only if they are both false:
Object is round OR object is blue
is true for blue balls, red balls, and blue cubes, but it is false for green pyramids.
In the case of SQL, you could use to to limit the number of rows SELECTED, or updated:
SELECT * FROM MyTable WHERE ID = 666 AND dateEntered > '2014-01-01' Would only return rows from me this year.
Make sense?
Those who fail to learn history are doomed to repeat it. --- George Santayana (December 16, 1863 – September 26, 1952)
Those who fail to clear history are doomed to explain it. --- OriginalGriff (February 24, 1959 – ∞)
|
|
|
|
|
I completely agree with what you are saying but here i mean to ask that
UPDATE [empdata] SET [name] = @p1, [id] = @p2, [salary] = @p3 WHERE (([name] = @p4) AND ([id] = @p5) AND ([salary] = @p6))
here what @p4, @p5 , @p6 would constitute?
what if i changed all the three values i.e. name, id and salary, then how would it know which row am i asking to change?
|
|
|
|
|
OK...combine the two comments I've given you and you have the answer.
Let's go back a stage or so, and look at a simple WHERE clause.
SELECT * FROM MyTable WHERE MyColumn=Value
The WHERE clause is "MyCoumn=Value" and it evaluates to a boolean: true or false
If it is true then the row is SELECTed (in this case, it could be UPDATEd in a different statement, such as yours) If it is false then it isn't.
The simple boolean test here is in three parts: "X = Y" where
X is saying which column of the row is to be compared
= is saying the the left and right sides must be the same
Y is the value to compare against the column
Working this back to your UPDATE, you have a three part WHERE clause, each part being of this form, and converted to a more complex test by AND to ensure that all three conditions must be true in order to update the row.
Each part is of the same form as my example above: "X = Y"
[name] = @P4
[id] = @P5
[salary] = @P6
If you change the names, it will check against different columns.
The part on the right is just the identifier of the parameter - it matches the name you use in your C# code to supply the actual value - look back at the first comment I made and it should be clear.
Those who fail to learn history are doomed to repeat it. --- George Santayana (December 16, 1863 – September 26, 1952)
Those who fail to clear history are doomed to explain it. --- OriginalGriff (February 24, 1959 – ∞)
|
|
|
|
|
I am creating a wrapper for a ThirdParty component. That component offers an event in case of an error:
public delegate bool ErrorOccurred(string id, string message); Of course, I'd prefer a custom exception, and wrote a class for that.
But there is a catch: I must return true to ThirdParty to indicate that we took care of the error, otherwise ThirdParty is likely to crash...
I subscribed to that event, and in the handler I try to do:
protected bool Session_ErrorOccurred(string id, string message)
{
throw new SASException(id, message);
return true;
}... but that doesn't work, of course.
I thought of throwing in a different thread (ThirdParty is multi-threaded anyway), but with
protected bool Session_ErrorOccurred(string id, string message)
{
Thread t = new Thread(() => ThrowSASException(id, message));
t.Start();
return true;
}
private void ThrowSASException(string id, string message)
{
throw new SASException(id, message);
}the exception gets swallowed somewhere in nirwana.
What do you suggest?
|
|
|
|
|
Perhaps your wrapper code could raise another event indicating an error occured and return true? That event's handler can choose to do whatever it wants (e.g. throw an exception). Not sure if this is a solution, but it may be worth trying.
/ravi
|
|
|
|
|
That's nasty!
Damn good idea, but nasty all the same...don't think there is a tidy solution for this, so it gets my vote.
Those who fail to learn history are doomed to repeat it. --- George Santayana (December 16, 1863 – September 26, 1952)
Those who fail to clear history are doomed to explain it. --- OriginalGriff (February 24, 1959 – ∞)
|
|
|
|
|
OriginalGriff wrote: That's nasty!
Agreed. 
OK then, I suggest not throwing an exception and instead simply having the consumer of the wrapper subscribe to the wrapper's OnError() handler.
/ravi
|
|
|
|
|
Trouble is...I'm not sure that even with an event the return will be executed at all times...if the exception isn't caught by anything, I think the app will crash. And I'm pretty sure that the exception will be immediately executed on the same thread and will pop the stack all the way back to the exception handler without ever hitting the instruction after the one raising the event...
This is nastier than I thought...I'm not sure there is any way to do this, short of a timer which is truly horrible!
Those who fail to learn history are doomed to repeat it. --- George Santayana (December 16, 1863 – September 26, 1952)
Those who fail to clear history are doomed to explain it. --- OriginalGriff (February 24, 1959 – ∞)
|
|
|
|
|
Griff is correct, this won't work – event handlers are called in the same thread, so if you re-fire an event and a delegate down the line throws an exception, your own handler code won't complete and you'll never return true to the library code.
This kind of thing is why it's usually considered a Bad Idea (TM) to throw exceptions in event handlers.
|
|
|
|
|
I agree with Ravi's solution, it's not perfect but I can't think of a better solution.
If you have the source code of the Third Party component or if there is a way to get them to change the component for you, you can ask them to include a parameter to indicate if the error was handled rather than returning true/false. Something like this:
protected void Session_ErrorOccurred(string id, string message, bool handled)
{
handled = true;
throw new SASException(id, message);
}
This signature still violates the Framework Design Guidelines for Events. It should ideally be like this:
protected void Session_ErrorOccurred(object sender, ErrorEventArgs args)
{
args.handled = true;
throw new SASException(args.id, args.message);
}
modified 5-Mar-14 11:14am.
|
|
|
|
|
This could get really ugly. You must allow the method to complete, so the library gets its result back, but you'd like the managed code to die with an exception as soon as an error occurs.
What is the exact conditions under which this delegate can get called? Is it that any call to third party library code could case OnError to be called, and then your code will continue afterwards?
I don't really see an alternative to
Library.CallToSomething();
if(ErrorWasRaised) throw new MyException("...");
Library.CallToSomethingElse();
if(ErrorWasRaised) throw new MyException("...");
... where ErrorWasRaised is instance state that's assigned from your OnError handler.
The only thing I can think of is to hide the ugliness in a utility method with a lambda:
void CallErrorHandlingLibraryCode(Action code) {
string errorId = null, errorMessage = null;
OnErrorHandler errorLogger = (id, message) => {
errorId = id; errorMessage = message;
return true;
};
Library.OnError += errorLogger;
try {
code();
if(null != errorId)
throw new SASException(errorId, errorMessage);
} finally {
Library.OnError -= errorLogger;
}
}
Then your calling code looks like
CallErrorHandlingLibraryCode(() => Library.CallToSomething());
CallErrorHandlingLibraryCode(() => Library.CallToSomethingElse());
... which is arguably better looking depending on your taste in lambdas.
|
|
|
|
|
If you are able to catch the exception and understand when it could occur then you shouldn't throw it. What you are trying to do is notify on the event. If you were throwing an exception then it would be fine to now return true. I know, seems counter-intuitive since all of the MS sample code is littered with throw exceptions every where but if you design with that in mind, your world can be a much better place. I have a lot of third-party libraries were I would kill for the ability to not have an exception.
|
|
|
|
|
The concept is wrong.
An exception must propagate the call stack of a thread.
Presumably what you want is the following
public void Call3rdparty()
{
try
{
Component.ErrorOccurred = MyErrorMethod;
Component.CallMethod()
}
catch(SASException e)
{
}
}
The problem with the above is that there is no assurance that MyErrorMethod will be called in the same thread as Call3rdparty
This means that the call semantics are implicitly asynchronous and it must be handled that way.
I believe there are some handy classes in C#/Net that allows one to create a async call but it isn't that hard to write one yourself using a monitor (or whatever it is called), timeout and appropriate attributes. If creating it yourself
- Create a thread to call the method -
- Set a monitor, with timeout
- If thread returns then set a flag, and set monitor
- If error method is called set attribute that indicates error
- Once monitor clears (excluding timeout) then check attribute for error, if there is error then throw exception
|
|
|
|
|
You are right. Thinking of the problem after my post, I came to a similar reasoning.
The error occurs somewhere in ThirdParty - though that process was somehow started through our code, it is quite un-coupled from it: we instantiate a class of their .Net SDK, call a Start method, which then communicates with ThirdParty server (that is on a different computer, could also be hosted in ThirdParty's cloud). On the server, the processing is started in a new thread, and the call to Start returns. When an error occurs in that new processing thread on the server, where could it be thrown to - it always travels the CallStack "upwards"? The Start method has long returned! Consequently, ThirdParty must use an event to communicate "downwards" to an "Observer". Hence I'll create a .Net event and raise that - then it is wrapped properly.
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
