|
When the keys are little sandboxes only pertaining to your application, why wouldn't you want that centralized there and why wouldn't they make it work more like that?
I don't really disagree and could see it going as you say... ditch the registry. I just don't get why IF the context is little sandboxed subsections of the registry instead of "you can edit anything or you can edit nothing".
|
|
|
|
|
That was the original idea: a single centralized depository for all config info. So it grew and grew, and bloated massively, slowed down everything and became a source of problems because there was no defined (or enforceable) separation of App A info from App B, much less system related info. It also added a route by which trojans could inject themselves into other apps very simply by modifying the config info.
So it became depreciated as a centralized store for new apps because of the problems it caused, and security features evolved as UAC was introduced which deliberately made it harder to use. Those continue to become more restrictive as the OS evolves - at some point it will probably become off limits to all non-admin apps (it's only not at present for backward compatibility with legacy applications).
If it had been designed as a sandbox environment in the first place, it could have been a really useful secure store - but it never was and that lack of foresight shows in the zombie version we currently have! 
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
"Common sense is so rare these days, it should be classified as a super power" - Random T-shirt
AntiTwitter: @DalekDave is now a follower!
modified 14-Feb-24 3:02am.
|
|
|
|
|
Richard Andrew x64 wrote: The registry key already exists when this code runs.
In that case, why are you using CreateSubKey rather than OpenSubKey?
CreateSubKey will attempt to create or open the key for write access. Since you're not writing to it, you should open it for read access instead.
Also note that the RegistryView is only required if your app is compiled as 32-bit and needs to access the 64-bit registry, or vice-versa.
Beyond that, you need to check the ACL on the specific registry key you're trying to open. It's possible that the system account has been denied (or has not been granted) access.
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
Richard Andrew x64 wrote: The registry key already exists when this code runs.
The exception doesn't complain about it not existing.
Richard Andrew x64 wrote: The code is running under the LocalSystem account as a Windows Service.
Which is, as documented, a limited account.
Richard Andrew x64 wrote: The initial open of the base key, HKLM apparently succeeds, but the BaseKey.Handle member is null immediately after the call to OpenBaseKey and before CreateSubKey.
That might be a hint 
Bastard Programmer from Hell 
"If you just follow the bacon Eddy, wherever it leads you, then you won't have to think about politics." -- Some Bell.
|
|
|
|
|
This comes from a comment in a prior post.
Global Usings – A Look at New Language Features in C# 10 | The .NET Tools Blog[^]
"With C# 10 you can use global using directives that the compiler will consider for the whole project"
To me this is just going to lead to similar problems as in C++ where one puts all of their includes in one include file and then uses that single include file everywhere.
It is indeed convenient.
But that doesn't mean that it should be used. Nor should it even exist.
When looking at a single source file one shouldn't need to be concerned about the impact that other files might have.
|
|
|
|
|
jschell wrote: When looking at a single source file one shouldn't need to be concerned about the impact that other files might have.
That's probably an unrealistic ideal. 
For example, the project file controls the C# language level, which can fundamentally change the meaning of the code in your C# files.
The project file determines which assemblies/packages you're referencing, which can also change the meaning and/or behaviour of the code.
Even going back to VB6, IIRC there was a project level "option base" setting which would change the default lower bound for array declarations which didn't explicitly specify one.
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
Yeah I'm not a fan. I'm not even all that wild about the namespace MyNameSpace; vs namespace MyNameSpace { } but mostly just because it breaks the code if someone happens to not be on more recent tooling.
|
|
|
|
|
Hi All,
I am trying to set up a 'framework' (not the right term) as I need to make use of using code sample below:
using System;
using Phidget22;
using Phidget22.Events;
I have got the example code set up and working example to compile but this is too complex for the function I am trying to do, I think as the correct steps have been followed to access the files. I think something has changed with the way to set up projects so they can see an external using file (or more likely I didn't do it properly!)
Glenn
|
|
|
|
|
Hi Glenn, it's not clear what your problem/question is.
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
To be perfectly honest neither am I, what I am trying to is be able from a Console app in C# 2022 to access the files used in the using statements above. I think I have it installed correctly as the example compiles and runs.
So is there a guide for doing this? I have had a Google and get not the right results. Last time I had to do something like this it was place the dll in the directory that doesn't seem to work...
|
|
|
|
|
Did you add a reference to the DLL in your C# project?
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
Hmm, it's a case of 'browse' to the DLL click OK and get a message it can't see it (but you can browse to it!)
|
|
|
|
|
|
|
Hi, I just upgraded a .NET Framework project to .NET 8.0 using the .NET Upgrade Assistant extension for Visual Studio from Microsoft.
I added a reference in the Dependencies folder to System.ServiceProcess.ServiceController.dll in the "C:\Program Files\dotnet\sdk\8.0.101" disk folder.
For some reason, it's unable to resolve types within that DLL (such as ServiceBase) and the reference in the Dependencies folder has a yellow triangle with an exclamation mark.
Is the project fubared because it was in-place upgraded, or is there something else wrong?
Why can't it resolve the types within that DLL, and why does the reference node have that yellow triangle icon? When I mouse over the icon, it doesn't display any tooltip to tell me what's wrong.
SOLUTION:
I added the file from the wrong folder. When I changed the reference to point to the same assembly in the "C:\Program Files\dotnet\sdk\8.0.101\runtimes\win\lib\net8.0\" folder, it resolved all the issues.
Thanks for reading my post.
The difficult we do right away...
...the impossible takes slightly longer.
modified 3-Feb-24 8:45am.
|
|
|
|
|
|
Hey, thanks. I didn't know it's available as a nuget package.
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
Many of what you'd probably think of as common to .NET dependencies are out there in various forms.
Dependency chaining has changed a little bit though. Mostly for the better.
For the most part, transients can now be consumed without a direct reference and very little effort.
If you have a nuget that then has <some nuget=""> as a dependency and you are already including that original nuget package, you do not generally also need a direct <some nuget=""> package reference in there... dotnet build will just figure that all out. (Assuming packagereference, but don't do packages.config anymore, ever)
|
|
|
|
|
the following is part for a c# code that I am working on:
powerShellScriptFilename = @"C:\Users\sherz\OneDrive\Downloads\enableTouchScreen.ps1";
ProcessStartInfo psi = new ProcessStartInfo
{
FileName = "powershell.exe",
Arguments = $"-NoProfile -ExecutionPolicy Bypass -Command \"{powerShellScriptFilename}\"",
RedirectStandardInput = false,
RedirectStandardOutput = false,
RedirectStandardError = false,
CreateNoWindow = true,
UseShellExecute = true,
Verb = "runas"
};
Process process = new Process { StartInfo = psi };
process.Start();
With this above the script run as expected.
The issue is if I move the script to the 'currentDirectory' (ie the Project debug folder) the script is not executed.
I already tried updating 'powerShellScriptFilename' with the corresponding absolute path, script still worth execute.
As the code works if I chose a different location, it seems to me that there may be some restriction or filepath length issue.
My aim is to try and avoid hard coding the filepath to the script and I would like my program to pick its current directory to find the script and run it.
How can I test/do that if it is not possible to do that from the debug folder?
still quite new to C# but keen to learn! 
modified 2-Feb-24 17:36pm.
|
|
|
|
|
How about:
powerShellScriptFilename = Path.GetFullPath("enableTouchScreen.ps1");
ProcessStartInfo psi = new ProcessStartInfo
{
FileName = "powershell.exe",
Arguments = $"-NoProfile -ExecutionPolicy Bypass -Command \"{powerShellScriptFilename}\"",
RedirectStandardInput = false,
RedirectStandardOutput = false,
RedirectStandardError = false,
CreateNoWindow = true,
UseShellExecute = true,
Verb = "runas"
};
Process process = Process.Start(psi);
Beyond that, we need to know what "not working" means. Provide the full details of the error you're getting. If it's a PowerShell error, then you may need to provide the relevant parts of your PowerShell script as well.
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
thank you. will try that when I get home.
"not working" means not working. not much more I can provide as I receive/can see no errors.
as you may have guessed from the fileames, the script enables my screentouch and that is how I know it was not successfull when I tried the current directory path.
with the original directory path, I would be able to enable my screentouch...
|
|
|
|
|
Sherzaad13291325 wrote: move to the script
Presumably you mean 'enableTouchScreen.ps1' for script.
Several obvious possibilities.
- You did not in fact move it.
- You did not update the path in your code.
- The path that you did update is wrong.
- You did not recompile the code after updating the path.
As one check for errors and probably one that I would always do in this type of code is to verify that the script (the path) exists before attempting to run it.
Additionally I might be inclined to check permissions also.
Doing both of the above is required if you intend for this code to eventually run in a Windows Service or a IIS App Pool.
|
|
|
|
|
Sherzaad13291325 wrote: the code works if I chose a different location, it seems to me that there may be some restriction or filepath length issue
My suspicion was right! turns out is WAS a filepath length. Converted the 'fullpath' to its 'shortpath' and code is now running as it did before 
modified 3-Feb-24 7:43am.
|
|
|
|
|
Based on the code you posted, no, it wasn't filepath length. The max length of a filepath is 260 characters. The max length of the Windows command line is 32K, and I seriously doubt your command line was that long.
|
|
|
|
|
You are right that the OP has not provided enough info for us to certainly conclude that it was a path length problem. But we certainly cannot reject it.
Then file name of his script, "enableTouchScreen.ps1", is 21 chars. He reports that when it is moved to "the 'currentDirectory' (ie the Project debug folder)", it won't work. If the length of 'cd' exceeds 238 chars, the full script file name would exceed 260 chars.
I certainly have seen project directory paths exceeding 238 chars; in my last job, the naming rules made it more or less the standard. What surprises me more is that the OP does not run into this problem with other files in his project as well. Maybe it is set up will all relative naming (and no tools expand it to an absolute path), or all files are located in other directories with shorter paths, or all the other tools are prepared for long file paths.
260 chars was the old DOS/FAT limit. As long as you use new Windows file system APIs that can handle longer NTFS paths, they can be up to approx. 32 Ki characters. The 260 length is not really in NTFS or Windows, but in individual programs using old APIs. These APIs have not been extended to long file names because lots of old program have allocated 261 chars for a file path buffer, frequently with no overrun check as that is the maximum length. If an old API used to return names limited by MAX_PATH after a minor update starts returning names of several hundred or thousand characters, then you would have a buffer overrun.
It is a pity that the myth of Windows being limited to 260 chars are still living. It makes lots of programmers continue making 261 char allocations, and the old APIs that should have been marked as deprecated 20 years ago and removed 10 years ago. Backwards compatibility is great (and Windows is far better at it than competing OSes!) but sometimes it goes too far.
Religious freedom is the freedom to say that two plus two make five.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
