|
Hi all,
I'm coding a small math library for an app we are developing. Its basically 3D math, Vectors, Matrix, etc.
I'm currently coding a struct Vector3D and I've overriden Equals(obj b) and was currently studying if I needed to override GetHashCode() or if the base implementation would work. A few simple tests proved that the default base implementation will not work as I'm getting same HashCodes for different Vectors and that's not good :p.
I'm not even sure I really need to implement a safe HashCode algorithm because I don't foresee the vector struct being used in such scenario, but still, I wouldn't like to leave something in my code that I know is not right. Never know how someone might consume the library in the future.
So my question is: Can somebody point me out how I should go about generating a unique int from 3 random doubles? My vector has obviously 3 coords, xyz and two vectors are the same only if all coords are the same which means that my GetHashCode() should generate a unique number for each combination of (x,y,z).
Any suggestions are welcome.
P.D. I'm not even sure if its possible to do. If I stop to think about it, you have a finite set of hashcodes you can generate (int.Min,int.Max) while the combination of diferent x,y,z coords is endless.  Am I misunderstanding the point of GetHashCode()?
-- modified at 19:13 Wednesday 9th May, 2007
|
|
|
|
|
Hi,
AFAIK has codes dont have to be unique (and as you pointed out, they
typically cant).
The one requirement is that the same object must return the same hash code
over and over. Basically it must always be true that
hash(obj1) != hash(obj2) ==> obj1 != obj2
You could satisfy this by having a constant hash (the hash based mechanism
does not provide any benefit then), or by having the hash equal to (or derived
from) just a few of the object's data bytes.
Of course if you expect to rely on effective hashing, then the hash values
should have a good distribution. So it is up to you to determine the balance
between costs and benefits.
|
|
|
|
|
Hi Luc, thanks for the reply.
I'm still missing something or I am completely confused and mistaken with the whole concept. Bear me out please:
The requirement of a hash code you are saying is:
if ob1 != obj2 ==> hash(obj1) != hash(obj2)
Ok I understand that, but it also has to comply with:
if obj1 == obj2 ==> hash(obj1) == hash(obj2)
I really dont see how you can manage this with a simple algorithm when it comes to Value types.
If we are talking about objects where reference equality is acceptable, then yes, it doesn't have to be unique with all possible cases, only with the existing object instances in memory...an easy hash that comes to mind in that case is returning the memoryaddress of the object (this will ensure the uniqueness of the hash and ensures that no other object can have the same hash while its in scope).
But with value types (and some reference values too I guess (string comes to mind)) this behaviour of the Hash algorithm isn't acceptable. For example, take this code:
public void Example(Vector v1)
{
bool same = areSame(v1, v1);
}
public bool whatEver(object o1, object o2)
{
return o1.GetHashCode()==o2.GetHashCode();
}
this code will return false if we implement a reference equality hash code which doesnt make sense, as its the same vector even if its boxed in two different memory addresses.
But how do we code a HashCode that will return you the same int for o1 and o2 but can assure you that no other existing vector in memory gives you that same code unless its also the exact same vector?
Obviously its impossible to get an algorithm that can give you a unique int for all possible existing vectors (we dont have enough ints to do that) so the solution must be to ensure that the hash works as expected with the existing set of vectors in memory.
Again, am I misunderstanding the meaning of GetHashCode() completely?
-- modified at 4:47 Thursday 10th May, 2007
|
|
|
|
|
Hi,
Sorry, my original reply was wrong. Of course there is only one requirement:
if obj1 == obj2 ==> hash(obj1) == hash(obj2)
The inverse is not necessary (and cannot be achieved in general, since one
cannot perform lossless compression from some amount of data towards a single
32-bit int).
I think the "remarks" in MSDN on Object.GetHashCode are clear.
It states:
"For best results, the hash code must be based on the value of an instance field or property"
and
"GetHashCode must always return the same value for a given instance of the object"
Now value types that get boxed twice produce two different objects,
their Hash Code does not have to be identical.
BTW I dont consider a Vector a value type.
Hope this helps.
|
|
|
|
|
Ok, thanks again for the reply.
I think I understand now. Anyhow, if you take a look at how MS generates hashcodes of ValueTypes like System.Drawing.Point you will notice that the same point boxed twice does produce the same hashcode even if they are not the same object, so I'd like to take that same approach. Does not seem that complicated anymore if we don't have to ensure uniqueness.
Just out of curiosity why wouldn't you implement a 3D vector as a value type as similar structs like System.Drawing.Point is? Am I missing some major drawbacks?
P.D. About MS comments on how GetHashCode() should behave I understand it perfectly but MS has one more recommendation and its that GetHashCode() should be compatible with .Equals(object o). Thats why the compiler flags you a warning that you should consider overriding GetHashCode() when you override .Equals(object o) (very normal situation when coding structs). That's what got me stumped as it's simply not possible to achieve 100% compatibility.
-- modified at 5:12 Thursday 10th May, 2007
|
|
|
|
|
Hi,
I was thinking of a Vector as it exists in Java (it is similar to C#'s ArrayList).
If your Vector is just a fixed and small number of coordinates, of course it
can be a value type.
|
|
|
|
|
hi all,
i want to detect all the pdf files link which are downloadable on webpage,
i need regular expression for that so that i can detect all the links.
can any one here to help me in that issue, i shall be thankful to my friend.
Regards,
Neo
|
|
|
|
|
Can't you just search for
[^\"\<\>\|\x0000\x0001\x0002\x0003\x0004\x0005\x0006\a\b\t\n\v\f\r\x000e\x000f\x0010\x0011\x0012\x0013\x0014\x0015\x0016\x0017\x0018\x0019\x001a\x001b\x001c\x001d\x001e\x001f\:\*\?\\\/]+\.pdf Basically what I've done here is used Reflector to look for all characters which .NET believes are invalid filename characters (the bit inside the [^...]+, the regex then looks for a group of one or more of these characters followed by a .pdf (\.pdf).
Have not tested it and do not accept resposability 
|
|
|
|
|
Thanks , But I am using Sockets and dont want to use Remoting.
So what If I put my assembly in GAC and App A and App B access it from there ? would not it be shared among these 2 applications ? and If I put singlton patteren on Class X to create only object then shoudl't App A and App B get same object ?
OR even if u put assembly in GAC , everny App creates its own instance of assembly ?
Thanks
|
|
|
|
|
Hi ,
I have question about .NET assemblies , say I have a Assembly 'X' and that conatins class 'xx' and my two different seperate applications uses this assembly , say Appl'A' and App'B' .
Then is it possible that I could share the object of class 'xx' in two applications ?
Like App'A' initiate object of class 'xx' and later on when App'B' comes, somehow I could tell that application that object of class 'xx' is already created and use that same object which App'A' has created.
Is it possible to do ?
I will really apprecaite if you could help me out here
Thanks,
ZINK
ZINK
|
|
|
|
|
You can accomplish this using .NET remoting: if AppA created the object, it would publish the object using .NET remoting's RemotingServices.Marshal(...) method. Then, AppB can connect to this same object using RemotingServices.Connect(...).
|
|
|
|
|
I have developed a glossary application whereby users can insert terms into the glossary. I have a function to check for duplicate word entries. Unfortunately there is a minor flaw here in the fact that a user can enter the same word with a leading space and it will be entered in the db as a new word. I.E. The Word and The Word[space] are to different entries. Is there a way to delete leading white spaces after/before a character string. Also I do not wish to delete words with spaces in between the words (i.e. The[space]Word). All suggestions welcome and future thanks to all that reply.
|
|
|
|
|
Have a look at Trim function an d its overloads of string class
|
|
|
|
|
Use the Trim() method of the string object, e.g.:
string textToEnter = " Citizen Kane";
textToEnter = textToEnter.Trim(); // it is now "Citizen Kane"
|
|
|
|
|
I'm running a service which puts out a log file as well as Debug.Trace.Writeline messages. When I'm on the server (as an administrator), run DBGVIEW.EXE, and then start the service I never see any trace messages.
Does anyone know what needs to be done so that trace messages can be seen??? Unfortunately, the problem I'm trying to debug on the server is the service either creating the new log or copying the old one to a backup directory....so I'm dying on my only alternative to viewing my tracing.
|
|
|
|
|
Hello,
I am trying to read a xml file using Memorystream, after 800 files i do get a exception
"Exception of type 'System.OutOfMemoryException' was thrown.
This is the code snippet, can any one pls suggest how to get rid of this?
<br />
public virtual bool TransformXML2( string strTempXml, ref string strTransformedXml )<br />
{<br />
bool bSuccess = true;<br />
if (m_strTransformFile.Length == 0)<br />
{<br />
PIUUtilitiesMgr.PIUMessage(PIUUtilitiesMgr.GetResourceString("MISSING_XSLT_FILE_NAME"), PIUUtilitiesMgr.GetResourceString("PERSONNEL_IMPORT_UTILITY"), MessageBoxButtons.OK);<br />
return false;<br />
}<br />
<br />
try<br />
{<br />
<br />
System.IO.MemoryStream streamTempData = new System.IO.MemoryStream();<br />
PIUUtilitiesMgr.FromStringToMemoryStream(streamTempData, strTempXml);<br />
string outText = PIUUtilitiesMgr.MemoryStreamToString(streamTempData);<br />
<br />
PIUUtilitiesMgr.WriteToFileForDebuggingPurpose("TempUserList.xml", outText, true);<br />
<br />
System.IO.MemoryStream streamTransformedData = new System.IO.MemoryStream();<br />
streamTempData.Position = 0;<br />
<br />
<br />
XmlReader reader = XmlReader.Create(streamTempData);<br />
XslCompiledTransform xslTransform = new XslCompiledTransform(true);<br />
<br />
XsltSettings settings = new XsltSettings(true, true);<br />
xslTransform.Load(m_strTransformFile, settings, new XmlUrlResolver());<br />
<br />
xslTransform.Transform(reader, null, streamTransformedData);<br />
<br />
streamTransformedData.Position = 0;<br />
<br />
strTransformedXml = PIUUtilitiesMgr.MemoryStreamToString(streamTransformedData);<br />
<br />
strTransformedXml = strTransformedXml.Remove(0, 1);<br />
streamTransformedData.Close();<br />
streamTransformedData.Dispose();<br />
streamTempData.Close();<br />
streamTempData.Dispose();<br />
<br />
<br />
<br />
<br />
if (piuInfo.DebugMode)<br />
{<br />
lock (this)<br />
{<br />
PIUUtilitiesMgr.WriteToFileForDe1buggingPurpose("TransformedUserList.xml", strTransformedXml, true);<br />
}<br />
}<br />
<br />
}<br />
<br />
|
|
|
|
|
You may want to watch your system with conditional breakpoints (like every 200 loops) and see if you are experiencing any memory leaks. In the earlier framework, the XML entities in the framework had a bad habit of leaking resources...ending in a memory not found condition.
In the 2.0 framework it was supposed to be fixed but I don't remember looking into it at all. So look into it. If you are experiencing issues then you've got to talk to your microsoft consulting rep and see what they can do to help you.
|
|
|
|
|
Hey all!
First off, a little background. I currently live in the UAE where the internet here is monitored and sites are blocked by our ISP's proxy server. Even applications like Skype are blocked here. So I decided that I want to try coding a little application that me and my friends would find useful for bypassing this proxy server. My plan is as follows:
(1) I already have a dedicated server in the US and I connect to it by remote desktop anytime I need to download anything that's blocked. I intend to use this server to re-route traffic to my computer.
(2) With the above in mind, I want to develop a client side proxy application. It would be running in the background, and the user would configure his IE or Firefox to the local computer's IP and a port number and start sending all requests through the client proxy app.
(3) The client proxy application would encrypt the request and send it to the server application sitting on my server in the US.
(4) The server would decrypt the request and fetch the pages and images of the sites that are blocked here in the UAE. After that, it would encrypt the result and send it back to the client which, in turn, would pass it off to the browser.
I hope all that made sense. Now, I am kinda new to network programming, so all I need is a point in the right direction to start me off. I'm also trying to get this to work for any application that connects to the net like Skype.
Thanks for any help guys! And feel free to suggest any other better ways I can go about solving my dillemma.
Cheers!
|
|
|
|
|
1) You want someone to help you gain access to sites banned by the govt?
2) Leaving the legality issue over here, why are you trying to write your own proxy application when you can let your server in US act as a proxy server and your browser/software use those proxy settings?
3) According to the laws of UAE, VOIP is illegal(I worked over there for a year). I last heard they are coming up with some kind of solution for the VOIP and you should be trying to use the legal way and not get into trouble.
This is the 3rd time this month a person who had created his/her account years ago(5 years in your case) comes up with a first posting which would have to be something stupid/illegal
|
|
|
|
|
Hello everyone,
I am using a DataGridView to display the entire data stored in a remote Database server. I am using the following code for populating the dataGridView controller.
OdbcAd = new System.Data.Odbc.OdbcDataAdapter("SELECT * FROM " + SelectedTableName + ";", OdbcCon);<br />
<br />
DataSet myds = new DataSet();<br />
<br />
OdbcAd.Fill(myds, SelectedTableName);<br />
<br />
dataGridView2.DataSource = myds.Tables[SelectedTableName];
I would like to have the Dataset continue to be available even after been disconnected from the remote database server.
Can someone tell me how this can be done?
Thank you very much and have a great day.
Khoramdin
|
|
|
|
|
What do you mean by saying available dataset? After disconnecting you dataset holds the data that was fetched when Fill method was called.
|
|
|
|
|
Hello Giorgi,
Thanx for the reply. Maybe my choice of words have created confussion. This is exactly what's happening.
I get connected to the remote database and populate the DataGridView with the data available in remote database. The moment I disconnect from the remote database server, I lose my entire data presented in DataGridView.
I would like to get connected to the remote database and populate the DataGridview with the data from the database server and then disconnect from the server but having the data still available in my Windows Application.
Thank you very much for your help, mate.
Khoramdin
|
|
|
|
|
Hello Giorgi,
I am so stupid.... lol
After my last posting I scanned my codes once again and realised this line of code which was written long ago is clearing the DataGridView.
It is good that I left the comment on the code! he hehe hehehe
this.dataGridView2.DataSource = null; // To Clear the DataGridView "dataGridView2".
Thanx for your help, anyways.
Khoramdin
|
|
|
|
|
Hi,
Anybody konws how to translate this VB code to C# code
Private Declare Sub PickIconDlg Lib "Shell32" Alias "#62" (ByVal hwndOwner As Long, _
ByVal lpstrFile As String, ByVal nMaxFile As Long, lpdwIconIndex As Long)
Thanks.
|
|
|
|
|
Have a look at www.pinvoke.net 
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
