Thread: "new" is strange

It finds the file, and it reads it correctly. But it still crashes. Oh and the release version crashes on dummy3 the debug version crashes on the next "new".

Originally Posted by The Wazaa

Hi, I have checked backwards. And as you said there was an error before that. Here is the area in which the error occures ( I think ):

Code:

fread(&temp, sizeof(int), 1, file);

This line should be OK, if temp is of type int.

Originally Posted by The Wazaa

Code:

	if (temp > 255 || temp < 1)
	{
		MessageBox(NULL, "Invalid mapname", "Error", MB_OK);
		return;
	}
	try
	{
		worldFileName = new char[temp];      // This line makes "new" go crazy...
	}
	catch (std::bad_alloc &)
	{
		MessageBox(NULL, "AIIEEE, out of memory", "Error", MB_OK);
	}

Unless your system is running out of memory, you're just seeing another symptom of your original problem.

I would bet that the real problem is in code executed even earlier than this code.

As a rough rule, there are only two things that can make operator new "go crazy": running out of memory (in which case an bad_alloc exception will be thrown and caught). If that's not the case, then your most likely offender is code that executed somewhere before the "crazy" behaviour became apparent.

Originally Posted by The Wazaa

Code:

	readStr(temp, worldFileName, file);		

	memset(worldFileName+temp, 0, strlen(worldFileName) - temp);
	MessageBox(NULL, worldFileName, "Info", MB_OK);

	WE_ERROR("Testing dummy3", EH_MBERROR, mainLog);
	int *dummy3 = new int;           // <---- This new works
	WE_ERROR("Dummy3 okay", EH_MBERROR, mainLog);

The fact that this "new works" means nothing. A problem has occurred earlier, and you need to find the cause of that.

As a rough rule, the symptoms of a problem cannot occur before the cause.

Originally Posted by The Wazaa

Code:

	fclose(file);

	TdsLoader *modelLoader = new TdsLoader;       // <--- This don't, it crashes

This is just a second symptom of your underlying problem. Which, I repeat, will be in code executed before this fragment.

Originally Posted by The Wazaa

This code is from a place where I try to read a file. I know that the file is correct since it works in debuggers. But when I run the exe, something strange seems to happen with all the strings and it looses them.

You are mistaken in your assessment that the "file is correct since it works in debuggers".

Code working correctly in debuggers, but not in production, is a symptom of a serious problem in your code (unless you have encountered a problem in the debugger or compiler: that is possible, but highly unlikely in your case).

The place you need to look is in code that is executed before the latest snippet you've given.

Since the problem could be anywhere (as grumpy said), you need to basically post the whole code for people to get a good look at all the things which could be going wrong.

If the project is still pretty small, then post it 'as is'.

If it's large, then you need to produce a smaller runnable example which still shows the problem.
a) make a backup / clone of the project
b) remove code which hasn't run yet.
c) experiment with commenting out code which has run, but which doesn't change the crash.

Hopefully, when you've done that, you'll have a much smaller (and easier for us to examine) example program to diagnose.

Often, just doing this will lead you to the answer you seek, as you eliminate possibilities.

Well, of some strange reason my strings get longer than I want, so I simply writes over the memory after it. Are you sure this causes the error? I thought that memset did not care about whether it writes outside the boundry of an array. And, if the length of the string would be as big as I wanted, it would not overwrite anything so I don't think that it is the problem. Its just an ugly way to solve that a lot of wierd characters seem to get into my string

Thanks for your reply

The Wazaa

Oh, I think I forgot to post readStr. And the characters that I speak about is in the string from the moment I create it. Check above for the readStr function.

But that is weird the way it runs from within the compiler. Maybe the working directories are set up wrong. With my compiler, my program always works when run inside the compiler, because the compiler sets the right working directories, but it usually crashes when I run the actual exectuable because it can't find the file that I'm trying to load.

> memset(models[i].fileName+temp, 0, strlen(models[i].fileName) - temp);
If you want a \0 terminator, just do
models[i].fileName[temp] = '\0';

Your strlen() assumes there is a \0 there to begin with, and if it's not there you end up with v.large values of strlen(), resulting in very large values of memory being cleared.

I could have sworn I asked you to post something which compiles and runs, yet you just posted random snippets all the same.
If you don't post full compilable code to such problems, expect half-assed answers and guesses in response.