Comments by wpte - CodeProject

wpte 12-Dec-14 8:00am View

Reason for my vote of 3 \n The idea is there, but it could be much simpler.
Right now you have an interface, 2 implementations and which wraps it all.

Why not 1 static class, static constructor which sets the corresponding filename in the static field to use by the pinvoke methods?

It'll be just 1 file, no wrappers, less code and faster! :D

Moreover, the exported methods you use dll-import on are static anyway, so any OO argument is invalid.

wpte 19-Sep-13 17:04pm View

I already use AppDomains extensively throughout my software.. Currently I only index the sub-directories once with reflection only load, so I can store the AssemblyName classes. When I'm done I throw away the AppDomain so no unnecessary memory is claimed. Since the construction of the cache only has to happen once, the resolving part is quite quick.

wpte 19-Sep-13 16:32pm View

I load plugins in my programs using Assembly.LoadFrom. Some of these plugins require other Assemblies (references) which are not located in the GAC. With the event I attempt to load these references by trying to find them in some sub directories.

wpte 19-Sep-13 15:50pm View

Allright. There's no way to check the correct version however. The only way to match the correct assembly is by matching strings. Bit unfortunate to be honest.

Still I'm convinced the event is called before .NET uses it's own resolver. I also get requests for System.something Assemblies which clearly need to be found in the GAC.

Also sorry for all the double posts, some code project maintenance was messing it up :(

wpte 19-Sep-13 15:15pm View

Deleted

Actually since .NET 4.0 the event is raised for every assembly.

"Beginning with the .NET Framework 4, the ResolveEventHandler event is raised for all assemblies, including resource assemblies. In earlier versions, the event was not raised for resource assemblies. If the operating system is localized, the handler might be called multiple times: once for each culture in the fallback chain."
http://msdn.microsoft.com/en-us/library/system.appdomain.assemblyresolve.aspx

If null is returned the .NET system will attempt to load the Assembly by itself, using the GAC for example.

I'm talking about embedded resources. I do get requests of these embedded resources. The request name is formatted something like: "Namespace.ResourceName.resources" instead of the usual AssemblyName.FullName format.

wpte 19-Sep-13 15:15pm View

Since .NET 4.0 the event is allways called before .NET tries to resolve the Assembly by itself using the GAC or local libraries in the same directory as the executable.

url: http://msdn.microsoft.com/en-us/library/system.appdomain.assemblyresolve(v=vs.100).aspx
"Beginning with the .NET Framework 4, the ResolveEventHandler event is raised for all assemblies, including resource assemblies. In earlier versions, the event was not raised for resource assemblies. If the operating system is localized, the handler might be called multiple times: once for each culture in the fallback chain."

I was talking about embedded resources. I get requests in the event with a format like: "namespace.resourcename.resources" not the usual format from AssemblyName.FullName

Now it's not that hard to find the accompanying Assembly which contains the embedded resource. As Collin Kleine suggests, returning an Assembly which contains the resource might be a good idea. I will check this tomorrow

wpte 19-Sep-13 15:13pm View

Deleted

Actually since .NET 4.0 the event is raised for every assembly.

"Beginning with the .NET Framework 4, the ResolveEventHandler event is raised for all assemblies, including resource assemblies. In earlier versions, the event was not raised for resource assemblies. If the operating system is localized, the handler might be called multiple times: once for each culture in the fallback chain."
http://msdn.microsoft.com/en-us/library/system.appdomain.assemblyresolve.aspx

If null is returned the .NET system will attempt to load the Assembly by itself, using the GAC for example.

I'm talking about embedded resources. I do get requests of these embedded resources. The request name is formatted something like: "Namespace.ResourceName.resources" instead of the usual AssemblyName.FullName format.

wpte 13-Mar-13 19:15pm View

Allright. I felt you came on a bit hard onto me, so I wrote a defensive repsonse. I understand managing a big site is difficult. I'm a moderator on a smaller site which can prove quite a bit of trouble now and then.

Thanks for explaining, and no hard feelings :)

wpte 13-Mar-13 18:01pm View

Well with custom I pretty much mean something like an alternate implementation, an override for example.
With these links I could probably cook something up, definitely a good read.
Thanks

wpte 13-Mar-13 17:54pm View

I find outrageous a big word to be quite honest. If the OP found a solution to his/hers problem it shouldn't be that weird to accept that as a solution when nothing else comes up.
I do agree with that my "solution" isn't really that much of a solution. The solution more or less is, that it's impossible. Point is that it worked for me and thus is a solution for me and thus I (apparently mistakenly) posted it as a solution.

Nevertheless, I apologize, next time I'll put it in a comment or improve my answer, whichever the most appropriate.

wpte 5-Jan-12 17:33pm View

code added... it works if I do it every second, but I want to have it continuously.
Don't ask me why I want that, I just do :)

The image is being edited with several filters with some OpenCL code. It's not an animation, it has to respond very quickly tho.

wpte 30-Sep-11 7:00am View

Ah right... silly me.
I forgot I declared buf within the same function :)
I need a coffee :p

wpte 30-Sep-11 6:52am View

Ok, I'm having problems with the delete []items part...

I have this in my header:
<pre lang="c#">
virtual ~buffer()
{
delete []items;
}
</pre>

but when I execute that using buf.~buffer(); glibc freaks out :p
I'm just storing constant integers in the buffer, not sure if I need to do that different.

glibc says: double free or corruption (fasttop): 0x0000000001f770c0

wpte 30-Sep-11 6:20am View

Makes sense, thanks :)

wpte 30-Sep-11 6:08am View

Ah thanks,

I've just implemented everything in the header file and it works great!
Thanks for helping me writing better c++ code :)
It's just, in all the tutorials I read so far on templates, they all implement them in the source file instead of the header file :)

wpte 30-Sep-11 5:56am View

I appriciate your help RKnGI ;)

wpte 30-Sep-11 5:52am View

I'm going to try this out... :)
Do you suggest that when I make other functions and methods that I should place them directly in the header file instead of the source file?

wpte 30-Sep-11 5:43am View

Well I've implemented that...
But it still doesn't explain how to use a parameterized constructor in a template...
I mean... how can I say that MyQueue should be holding at most 10 T's?

Comments by wpte (Top 18 by date)