Comments by Anthony Mushrow (Top 48 by date)

If you have a question regarding your own code and you can't find a suitable answer, then you should post a new question. Do not post you question in response to somebody else's.
However, your problem is that the contents of the getTableData function has been commented out and it no longer returns anything.

If you're getting an Internal Server Error then it's likely an exception is being thrown on the server. If your using Visual Studio you can run the website with the debugger attached, you can either set a breakpoint and step through to make sure everything is all right, or if you're getting a First Chance Exception when calling your GetProblems API you can halt the execution of the program by enabling the exception by going to Debug->Exceptions.

I think I've figured this one, the multiplication uses inline asm to do the actual work in release we get return value optimization, the result of which means that the asm no longer stores the result in the correct place, or rather the compiler moves the goalposts and now I have no idea where the value should be stored.

It's not the actual time that an object is destructed that I'm worried about, It's that objects are destructed before I can use them. Surely something is flawed here, how can I possibly make use of the result of a matrix multiplication if the object is destroyed before I get to see it?

And why would my second thing object be destroyed, it not a few seconds later or when the object is finally destroyed it's immediate. After making the second Thing object if I try to use it the application will crash because I'm trying to reference on object that has been destroyed. These issues aren't present in a plain C# project.

I was hoping to avoid making some defines for it, at least there's not too much around so making them all use the defines shouldn't take too long.

Looking at the brighter side if I ever need to support any different syntax I can just make a new define.

Yeah I made a mistake, it's difficult to tell where everything is when nothing is indented. Like Philippe Mori said, you need to uncomment line 125 and use x.evaluate

You should translate and then scale, otherwise the translation will be scaled as well.

If you could not bring back a selection of ancient questions to spam links to a website it would be much appreciated.

Would you be able to elaborate on what is wrong with the code? I assume the OP already knows that there is something wrong so this answer doesn't really help.

Why would std::vector be better? The only thing you'd really get our of it is not having to remember the delete, plus some overhead for going through the container. If the size of the array isn't going to change after it was created I don't really see any advantage.

It seems you'll need to have access to the existing instance of the form, instead of using a new one. How you get hold of the instance entirely depends on how your application is set up, so I can't really help there.

I see the difference now. It's still a bad idea to use exceptions to do this sort of thing. After a quick check OpenSubKey will just return null if the key is not found, so you'd never even get an exception and try the other key. The only exception you're likely to get is for not having permission to edit the registry.
I've updated my answer with a modified version of your code that should actually work.

There are an infinite number of points on a line so you're going to need to clarify what points you're actually after. Do you mean points on a grid that the line passes directly through?

Yeah, it's pretty weird. I'm a little surprised the reference device failed though. My best guess is that the problem isn't with anything your doing, especially if it still actually works with the same settings.

It might be worth trying it on a different machine

Well actually, you could use the Win32 SetParent function to make the panel the parent of the other form. But err, not recommended.

Which I also answered, the code in the lib file is linked into your application, even if the lib file is only a thin wrapper for calling the DirectX runtime DLL's so you don't need it any more.

No, but it did answer the question directly. You don't need to have the .lib file on the target machine and older versions of DirectX will work on a machine with the latest runtime installed.

What actually happens is that each thread is given some time on the processor before it is switched out with another process so the threads aren't actually running in parallel, they are just interleaved with each other. Also switching threads a lot can hamper performance (assuming all of the threads are actually busy, rather than waiting for events or data). Think of a computer that locks up because a single greedy process is using all of the available cpu time, with other threads getting very little time to do anything.

I can't explain why 10 busy threads would be able to handle more data, unless the work that they do involves waiting on other hardware (reading from disk, network communications etc.) in which case the optimum number of threads to use is going to be a little more difficult to find. For certain though you don't want to just keep creating threads forever, you might be better off working something out with a ThreadPool, which will manage the threads for you.

You can make as many threads as you like, but if you have a quad core cpu then yes, only 4 of your threads will ever be running at exactly the same time.

Environment.ProcessorCount should tell you the number of logical processors / cores that are on the current machine, you can use that to determine how many threads you should create

Well, the window a user is working on is always on top and as such is the foreground window. From MSDN Retrieves a handle to the foreground window (the window with which the user is currently working).
What's your definition of the Active window? Did GetForegroundWindow not return what you had expected?

The error code says that you made an invalid call to the function. Could you post a code snippet showing how and where you reset the device?

Answer updated.

I shall repeat myself. I used WaitForMultipleObjects because I had a breakpoint before the application returned so that I could see the output in the window before it closed. If the breakpoint was hit before the threads had printed out their message then visual studio would have halted the threads as well, meaning I may not have been able to see all of the output.

Just because I use WaitForMultipleObjects it does not mean that the threads will not run in parallel, I'm not entirely certain what you think it does but what it really does is block the calling thread until X number of handles have returned (which in this case is when all of the threads have finished). It has absolutely no effect on the threads themselves.

Well, my example was to show that CreateThread works just fine and that 5 threads are created and run. The problem the OP seems to have is that only the first call to CreateThread works. And what's pointless about my example (apart from the output)?

First of that code was a quick example by me, not the OP. I added the WaitForMultipleObjects because I had a breakpoint on return 0; so I could see the output, and if it was hit before the threads had finished then visual studio would have paused their execution too.

It's true that starting a thread that calls a one line function is probably going to be the same as just calling that one line in sequence but not necessarily so. Also, the output is not serialized back to a one-by-one situation by use of WaitForMultipleObjects, all that does is block the main thread until all of the other threads have finished. If you add a Sleep(5000) to the threads function then you'll only have to wait for 5 seconds, not 25 because the threads are executing in parallel.

I just did a quick check with CreateThread myself and everything works as expected, is your setup significantly different to this:

#include <windows.h>
DWORD WINAPI t_func(PVOID args){
printf("Threaded Write\n"); return 0;
}
int _tmain(int argc, _TCHAR* argv[]){
HANDLE Threads[5];
printf("Bacon\n");
for(int i=0; i<5; i++)
Threads[i] = CreateThread(NULL, 0, t_func, NULL, 0, NULL);
WaitForMultipleObjects(5, Threads, true, INFINITE);
return 0;
}
Output:
Bacon
Threaded Write
Threaded Write
Threaded Write
Threaded Write
Threaded Write

Sorry, I meant that for a program using DI to get input from the keyboard the SendInput API might not have any effect. The sent keypresses might not show up when the device is polled.

DirectInput part of DirectX, allows exclusive access to input devices meaning that there isn't any (or at least very little) handling done by windows between the program using DirectInput and the hardware itself. Games use DirectInput to get user input from game controllers and keyboards, and they're mostly the only things that do.

I don't think SendInput works on that low a level. So far as I'm aware it can't fool DirectInput or other similar input libraries. It's not a problem in this case but maybe something to be aware of.

Please don't add answers when you want to leave a comment to somebody else's answer, there is a small link at the bottom right of each answer labelled 'Add Comment' that lets you add a comment to an existing answer.

You should use the <pre> tag when posting large code segments as to maintain the formatting as well as to get some fancy colourising.

Sorry, forgot to encode my < > for the includes in my update. You need to add #include <iomanip>

If you're using cout then you need to use setprecision to specify the number of decimal places you want.
float num = 1.23456f;
cout << "My Number: " << setprecision(1) << num;

Actually on my device you have the choice of 2G only, 3G only, or just let the phone use whichever network has the strongest signal. So presumably there must be something somewhere, although it may be device specific.

Good catch, I was partially right at least.

The assert should be able to tell you what's going wrong, that's what it's there for. Nobody can really help with the information you've given, the only things I can think of that could cause some kind of assert is that either the surface has already been released, it was never created in the first place (the function failed) or it's locked or something. Read the assert, look it up using google, you should be able to find your answer.

Would you be able to tell us what the error is perhaps?

Hmm, so adding more threads is kinda like using too much memory and causing paging, or setting the max resolution in a game with the highest AA when your card doesn't have enough RAM to handle it all.

I always figured that threads would be handled better somehow. But still, like a good developer I am limiting the number of threads created based on the number cores to run them.

If a user knows enough to go into the registry and start deleting keys, then you could assume they'd know enough to go into the windows directory and start deleting suspicious files.

You need to be careful about posting your email on the internet, or you inbox will be filled with thousands of spam emails.

After more searching is seems you're partially right. There is a bug in PHP on windows where it doesn't parse the friendly name thing correctly, it's actually mentioned at the bottom of this page:
http://php.net/manual/en/function.mail.php

Would have been nice to know sooner.

I believe Windows 2008 would be Windows Server 2008