Re: Is there a difference? - C# Discussion Boards

RNEELY2-Mar-04 11:58

RNEELY

2-Mar-04 11:58

No.

Sincerely,
-Ron

Heath Stewart2-Mar-04 12:26

2-Mar-04 12:26

To explain why RNEELY simple answered "no", it's because the compiler assumes that any un-qualified calls use the this reference anyway. It's an implicit object. When it compiles, the exact same Intermediate Language (IL, the language embedded in modules of which an assembly is partly comprised) is produced. In both cases, the optimized body of each method would look something like this:

ldfld int32 MyClass::i1
ldfld int32 MyClass::i2
add
ret

Microsoft MVP, Visual C#
My Articles

Colin Angus Mackay2-Mar-04 12:28

Colin Angus Mackay

2-Mar-04 12:28

Okay... Here's a question for you...

[Edit]
Ooops! The original version of this program had an error in it that I didn't intend. Here is the correction, the multiple choice answers were as before.
[/Edit]

What is the output of this program:

01: class App
02: {
03:     int i1 = 0;
04:     int i2 = 0;
05: 
06:     public void Method1()
07:     {
08:         i1 = 2;
09:         i2 = 3;
10:     }
11: 
12:     public void Method2(int i2)
13:     {
14:         i1 = i2;
15:     }
16: 
17:     public static void Main()
18:     {
19:         App a = new App();
20:         a.Go();
21:     }
22:
23:     public void Go()
24:     {
25:         Console.WriteLine("i1 = {0}, i2 = {1}", i1, i2);
26:         Method1();
27:         Console.WriteLine("i1 = {0}, i2 = {1}", i1, i2);
28:         Method2(5);
29:         Console.WriteLine("i1 = {0}, i2 = {1}", i1, i2);
30:     }
31: }

Is it:
a)
i1 = 0, i2 = 0
i1 = 2, i2 = 3
i1 = 3, i2 = 3

b)
i1 = 0, i2 = 0
i1 = 2, i2 = 3
i1 = 5, i2 = 3

c)
i1 = 0, i2 = 0
i1 = 2, i2 = 3
i1 = 5, i2 = 5

d)
None of the above - it generates a compiler error on line 12

EuroCPian Spring 2004 Get Together[^]
"You can have everything in life you want if you will just help enough other people get what they want." --Zig Ziglar

boogs2-Mar-04 13:08

boogs

2-Mar-04 13:08

boogs guesses c) -> arguments take precedence over members.
and no, i didn't test it - that would be cheating Wink | ;)

boogs2-Mar-04 13:13

boogs

2-Mar-04 13:13

d0h, it was b). for the same reason.

Dave Kreskowiak2-Mar-04 13:33

Dave Kreskowiak

2-Mar-04 13:33

E) None of the above

The signature for Main is an invalid one for an entry point. Also, changing it to static would preclude Main from accessing the public members of the App class without first creating an instance of the class. Laugh | :laugh:

RageInTheMachine9532

Colin Angus Mackay2-Mar-04 13:38

Colin Angus Mackay

2-Mar-04 13:38

Okay - let's assume I didn't make that error... D'Oh! | :doh:

EuroCPian Spring 2004 Get Together[^]
"You can have everything in life you want if you will just help enough other people get what they want." --Zig Ziglar

Meysam Mahfouzi2-Mar-04 17:55

Meysam Mahfouzi

2-Mar-04 17:55

You know, this keyword is mostly used whenever in a function, the passed in parameter has a name the same as the name of a field:

private string name;
public void SetName(string name)
{
       this.name = name;//this.name refers to private field
}

Don't forget, that's

Persian Gulf

not Arabian gulf!

Murphy:
Click Here![^]

I'm thirsty like sun, more landless than wind...

sps-itsec463-Mar-04 8:48

sps-itsec46

3-Mar-04 8:48

Thanks for your comment... what you say makes totally sense! Smile | :)

OT: You should check the link to the murphy page in your signature it's not "http://www.thecodeproject.com/..." but "http://www.codeproject.com/...". Glad I could help you too Big Grin | :-D

Regards, mYkel

HELP with Code Access Security!

Meysam Mahfouzi3-Mar-04 17:27

Meysam Mahfouzi

3-Mar-04 17:27

It's really pleasure that it made sense!;)
BTW, thanks for notification about signature

Don't forget, that's

Persian Gulf

not Arabian gulf!

Murphy:
Click Here![^]

I'm thirsty like sun, more landless than wind...

Throckmorton2-Mar-04 11:10

2-Mar-04 11:10

Hello,

I am wondering if there is a way to programatically find a code group in the configuration wizard?

Basically, I have a c# app (local) that calls c# apps (remote share). If I try to call an app without the proper permissions set up, the CALLED app seems to throw a security exception. Since I have no idea how to trap an exception in localApp that is thrown by calledApp, I want to try something else:

What I want to do is this:
1) Before calling remote app, somehow query the Machine Code Groups in .NET Framework Configuration (programatically).
2) If the required code group does not exist, create it.
3) Finally, call the remote app.

Is this possible?

Thanks,
Ian

Heath Stewart2-Mar-04 12:19

2-Mar-04 12:19

Yes, you can enumerate the code groups. You can enumerate the SecurityManager.PolicyHierarchy to get the right policy (like Machine), then enumerate the code groups underneath that. I do it in are installer for our touchless-deployment app (well, not completely touchless since a one-time installation is required). You can further add code groups using this mechanism, but you should at least prompt the user. Companies and savvy users don't like you messing with security unless it's clear that you are and why you are messing with security.

Really, though - if both of these applications are managed applications, you should consider another approach. If you don't mind that one "process" runs within the first, you can create a new managed AppDomain and load the assembly that contains the entry point Main from the second application using AppDomain.ExecuteAssembly. You can either pass your current AppDomain's evidence, or gather host and assembly evidence for the new AppDomain and pass that to ExecuteAssembly. If the executable does not have the necessary permissions, a SecurityException is thrown at that time.

Microsoft MVP, Visual C#
My Articles

Throckmorton2-Mar-04 13:58

2-Mar-04 13:58

Thanks Heath -

I hate to ask but given that I am a total noob at this - can you provide (or point me to) a more specific example of enumerating the PolicyHierarchy and searching for specific groups? I am unable to actually resolve specific group names or at least cannot figure out how to.

Right now I have:
IEnumerator levels = SecurityManager.PolicyHierarchy();
while(levels.MoveNext())
{
PolicyLevel pl = (PolicyLevel)levels.Current;
Console.WriteLine("Policy Level: {0}", pl.Label);
}

Which prints:
Policy Level: Enterprise
Policy Level: Machine
Policy Level: User

I see a method in there called ResolveMatchingCodeGroups() but it wants an "evidence" and no matter what I try in the Evidence object it always returns the same things, something like:
All_Code
All_Code
All_Code

I want to check if "MyCustomCodeGroup" exists. It is located here:

- Runtime Security Policy
    - Machine
        - Code Groups
            - All_Code
                - LocalIntranet_Zone
                    - MyCustomCodeGroup

The membership condition type is "URL" and the URL is "file://MyServer/*".

The permission set on this group is FullTrust.

As for the other part of your reply - The appdomain/assembly stuff is complete greek to me. I will have to learn more about it. However I would say that the local exe calls remote exe's. The remote exe's can be managed code but also they can be .cmd/.bat files or even .msi, or .exe (C++ apps).

Thanks for your help,

Ian

Throckmorton2-Mar-04 17:23

2-Mar-04 17:23

Ahh I figured it out finally Smile | :)

Heath Stewart3-Mar-04 3:27

3-Mar-04 3:27

The important thing is not the name of the code group, but that the permission set assigned to the code group is sufficient to run your code. What if someone changed the name of the code group or created their own? It's the CodeGroup.PolicyStatement - or rather the PermissionSet that the PolicyStatement contains - that is important.

Instead of checking that, though - as there would be much more to do - you could get the evidence for your new assembly and compare that against the CodeGroup.MembershipCondition to see if they match.

As far as launching any executables besides .NET assemblies, code access security doesn't really matter except for the ability to access that file on whatever media it resides. Your application that launches that application must have permissions to access that resource (read access, in order to launch it). It is your application that must have it, so if you want to determine whether or not you can before you launch it, you should create a FileIOPermission with the path to the resource and in a try/catch call Assert. If no exception is thrown, you can go ahead and launch the executable. Catch any SecurityException that is thrown and display and error, log it, or whatever.

Microsoft MVP, Visual C#
My Articles

Throckmorton3-Mar-04 9:19

3-Mar-04 9:19

Thanks again,

I think I have this working to my satisfaction (for the time being at least - till I figure out more of how all this works).

What I do is enumerate down to the groups below LocalIntranet_Zone and look for MyCustomAccess. Then I also check that it is type URL and points to the appropriate share location. I also check that it has FullTrust permission set.

Here's my next problem:
If I do NOT find the proper code group with proper permissions, I want to add one. I have seen many examples in various places. Every example creates a new code group at the same level as LocalIntranet_Zone (siblings(?)). I want to create a new group that is a child of LocalIntranet_Zone. While I haven't keyed in the examples, I assume they work. The issue I have is when I call SecurityManager.SavePolicy() - it does not actually get saved, even though the group was added or at least seems to have been added.

Here is my code: can you tell me what is wrong? Why doesn't it save?

		<br />
static bool addCodeGroup() <br />
{<br />
		<br />
	// Enumerate all the policies<br />
	IEnumerator policyEnumerator = SecurityManager.PolicyHierarchy();<br />
<br />
	// Move through policies until we hit Machine<br />
	while (policyEnumerator.MoveNext()) <br />
	{	<br />
			<br />
		PolicyLevel machineLevel = (PolicyLevel)policyEnumerator.Current;<br />
<br />
		// If we are in the Machine Security Policy...<br />
		if (machineLevel.Label == "Machine") <br />
		{<br />
			// Build an enumerator to move through the machine level groups<br />
			IEnumerator ie_MachineLevelGroups = machineLevel.RootCodeGroup.Children.GetEnumerator();<br />
					<br />
			while (ie_MachineLevelGroups.MoveNext()) <br />
			{<br />
						<br />
				CodeGroup machineLevelGroup = (CodeGroup)ie_MachineLevelGroups.Current;<br />
<br />
				// if we are in the LocalIntranet_Zone Code Group...<br />
				if (machineLevelGroup.Name == "LocalIntranet_Zone") <br />
				{<br />
<br />
					PermissionSet permSet1 = new NamedPermissionSet("FullTrust");<br />
					IMembershipCondition membership1 = new UrlMembershipCondition("file://wstltest/*");<br />
					PolicyStatement policy1 = new PolicyStatement(permSet1);<br />
					CodeGroup codeGroup1 = new UnionCodeGroup(membership1, policy1);<br />
					codeGroup1.Name = "WSTLTEST_ACCESS";<br />
					machineLevelGroup.AddChild(codeGroup1);<br />
					SecurityManager.SavePolicy();<br />
									<br />
<br />
						} // end if (machineLevelGroup.Name == "LocalIntranet_Zone") <br />
<br />
					} // end while (ie_MachineLevelGroups.MoveNext()) <br />
<br />
								} // end if (machineLevel.Label == "Machine") <br />
<br />
	<br />
			} // end while (policyEnumerator.MoveNext()) <br />
<br />
<br />
			return true;<br />
<br />
		}

Heath Stewart3-Mar-04 9:27

3-Mar-04 9:27

While the reference is correct for the machinePolicy object you use, try SecurityManager.SavePolicyLevel(machinePolicy);. I use this in our installer with no problems.

Microsoft MVP, Visual C#
My Articles

Throckmorton3-Mar-04 10:11

3-Mar-04 10:11

Thanks Heath, but still not working -

Where would I add SecurityManager.SavePolicyLevel(machineLevel) call?

I tried adding it just before SavePolicy() call...
Just after SavePolicy() call...
In place of SavePolicy() call...

Still doesn't save the machine policy.

Heath Stewart3-Mar-04 10:16

problems in running app at win startup

3-Mar-04 10:16

Compare it with this...which works:

// Create the policy statement.
NamedPermissionSet namedPS = new NamedPermissionSet("FullTrust");
PolicyStatement ps = new PolicyStatement(namedPS);
ps.Attributes = PolicyStatementAttribute.Exclusive;
 
// Create the membership condition.
string url = "http://www.dummycorp.com";
UrlMembershipCondition condition = new UrlMembershipCondition(url);
 
// Create the code group.
UnionCodeGroup proplanner = new UnionCodeGroup(condition, ps);
proplanner.Name = "MyCodeGroup";
proplanner.Description = "Allows fully-trusted access to the URL.";
 
// Find the machine policy level.
PolicyLevel machine = null;
IEnumerator ie = SecurityManager.PolicyHierarchy();
while (ie.MoveNext())
  if (String.Compare(((PolicyLevel)ie.Current).Label, "Machine", true) == 0)
    machine = (PolicyLevel)ie.Current;
 
// If the machine policy level was not found, throw an exception.
if (machine == null)
  throw new InstallException("Unable to find the Machine policy level.");
 
// Enumerate the root code group children and
// determine if the Proplanner group already exists.
bool exists = false;
foreach (CodeGroup group in machine.RootCodeGroup.Children)
{
  if (proplanner.Equals(group))
  {
    exists = true;
    break;
  }
}
 
// If the code group doesn't exist, add it.
if (!exists)
{
  // Add our code group to the root code group.
  machine.RootCodeGroup.AddChild(proplanner);
  SecurityManager.SavePolicyLevel(machine);
}

Microsoft MVP, Visual C#
My Articles

visiontec2-Mar-04 10:22

visiontec

2-Mar-04 10:22

Hi there

I am having problems with loading my app at windows startup.

When i load my app by clicking on it, it works fine
but when i load it on windows startup, by adding a registry key in
HKLM\Software\Current Machine\Microsoft\Windows\Current Version\Run
it gives errors, the errors are of my files that i load when my app starts.

i have an xml file that i load when my app starts
that xml file has the location of my database file.

for now i store my db file where my app's executable is.

The code for my class :

using System;
using System.Xml;
using System.Xml.XPath;

namespace Scheduler
{
   public class SettingsXml
   {
      XmlDocument doc = new XmlDocument();

      public SettingsXml()
      {
         doc.Load("scheduler.xml"); // this means that this file
                                    // is where the app's .exe is
      }

      public string DBPath()
      {
         string xstr = "/Scheduler/Configurations/DbPath";
         XmlNode xnode = doc.SelectSingleNode(xstr);
         return xnode.InnerText;
      }
   }
}

when the app loads at win startup it looks for my "scheduler.xml" file
in "c:\scheduler.xml" rather then the place i have told it to!

I fixed the "scheduler.xml" file's location problem by doing this:

doc.Load(System.Windows.Forms.Application.StartupPath + "/scheduler.xml");

It worked.
But still in my xml file the path of my db file is "scheduler.mdb"
now what do i do for this?

cause when i distribute my app it has to have some default place to place my db file.

What is the problem here?

VisionTec

Re: problems in running app at win startup

Dave Kreskowiak2-Mar-04 11:12

Dave Kreskowiak

2-Mar-04 11:12

visiontec wrote:
doc.Load("scheduler.xml"); // this means that this file // is where the app's .exe is

Unfortunately, no it doesn't. The file is tried in the current directory. The only reason why that works in your testing is because you launched the EXE from the folder by double clicking on it, thereby the current directory is the folder you double-clicked in.

When you launched the EXE from the Run key, the current directory is C:\, or one of the Windows directories, depending on what's going on. Your app is being launched with the full path to the EXE, but that doesn't make it the current directory.

To get around this problem, best-practice would be to code full path names into your app, but NOT hard-coded paths. Use configurable paths like:

StreamWriter sw = New StreamWriter(datapath + "\scheduler.txt");

You could also set the current directory when your app starts by using

Environment.CurrentDirectory = datapath;

But, best practice is to assume nothing! Don't assume that because you set the CurrentDirectory, that it is going to stay there!

EDIT:
Whoops! I supplied the code examples in VB, not C#. Rewrote for C# examples.

RageInTheMachine9532

Client size of an control ...

Andres Coder2-Mar-04 7:49

Andres Coder

2-Mar-04 7:49

Imagine this - I have two controls, one is simple usercontrol and the other is derived from System.Windows.Forms.Form.

Does any these controls have client area size different than the usual size of the control (Height and Width properties) ? Like - a form has a thick border around it, is the width of the border added to the Height property of the control ? Or when a usercontrol has a 3d border ?

I've heared a lot about nonclient area and stuff like that, where could I get more information about that ?

Regards, Desmond

Re: Client size of an control ...

Heath Stewart2-Mar-04 9:24