Comments by Niklas L - CodeProject

Niklas L 18-Sep-12 4:24am View

Good indeed

Niklas L 18-Sep-12 4:22am View

Ah! Didn't know that! Good info.

Niklas L 17-Aug-12 5:04am View

Should you be using 'numItems'? Looks like you should be using 'i' to index your arrays (or whatever they are)
Are you sure the number of items in FirstName equals the number of items in your other collections?

Niklas L 15-Jun-12 4:09am View

That's not necessary if you set the buddy control manually.

Niklas L 14-Jun-12 5:38am View

You need to review your code and clean it up. For example:
//excercises_LL.removeAllViews();
excercises_LL.setWeightSum(1.0F);
excercises_LL.setVisibility(0);

Why have you removed removeAllViews()?
Why do you set a weight sum when you don't know what that sum is?
Why do you use numeric values, when you should use named constants like View.VISIBLE, View.HIDDEN and View.GONE?

You are not making your life easy.

Now, all you need to do is to set the texts of your TextViews again using setText(). To do this you need to make your content_LL identifiable. Use the id of the table record (_id) and set it as the views id. Then you can find it again using excerise_LL.findViewById(). If you don't have a numeric key, you could make use of the setTag() method to make it identifiable.

Niklas L 13-Jun-12 14:04pm View

Ah, beaten. I send a five in your general direction.

Niklas L 13-Jun-12 4:37am View

Have a 5

Niklas L 12-Jun-12 16:00pm View

thank you

Niklas L 26-Apr-12 5:34am View

It's not "classless". It should look like:
class HelloOpenGLES10SurfaceView extends GLSurfaceView {
...
public HelloOpenGLES10SurfaceView(Context context){
...
} // end construtor
@Override
public boolean onTouchEvent(MotionEvent e) {
}
} // End class

Niklas L 26-Apr-12 5:11am View

In that case, place it outside of HelloOpenGLES10SurfaceView's constructor.

Niklas L 22-Apr-12 4:40am View

What I meant was that if you only had written the additional EDIT part, it would be a five alone. We do not disagree at all.

Niklas L 19-Apr-12 4:15am View

Love your rants ;) Welcome back!

Niklas L 19-Apr-12 4:13am View

5 for the EDIT part alone. The rest is bonus.

Niklas L 19-Apr-12 4:02am View

A good solution, but it leaves an unnecessarily large memory footprint, which you will suffer from when copying instances around. A single double (or float) would probably do. 5'ed anyway, for making it an OO solution.

Niklas L 17-Apr-12 7:40am View

Oh, that's a bit more specific ;) What media types will you be streaming? There seem to be a few libraries around, specializing in different types. Maybe PJSIP could work for you (http://www.pjsip.org). It was originally designed for audio only, but Version 2 branch looks promising for video.

Niklas L 27-Mar-12 6:04am View

Have you checked of it's not just about adding an extra flag to the intent?

Niklas L 6-Dec-11 5:24am View

Thank you! I cannot understand why people insist on using new (not to mention malloc!) instead of standard library classes, when learning C++. I suppose the literature is of poor quality in general.

Niklas L 18-Nov-11 6:07am View

Looks like you are missing to include a library file for JNI in your project.

Niklas L 10-Nov-11 7:33am View

How do you give the path to the mdb-file? Something there maybe (?)

Niklas L 20-Oct-11 4:51am View

I can only think of three possible reasons for this.
1. There is a typo in the file name/path
2. The jpeg image is using some format extension not supported by default by windows.
3. The program is not authorized to change the wallpaper.

1. Easy to check
2. ~~Does it open in Paint? (Don't use a fancy image program to test if it opens)~~ If it doesn't, try another image to see that the API works. Edit: The correct way to test this is of course to set the wallpaper manually as a .. ehum.. wallpaper.
3. I currently have no solution for this.

Niklas L 17-Oct-11 16:33pm View

Of course you would know where it is. Did you even try
master[index1].pChild[index2]?

Niklas L 14-Oct-11 7:43am View

Deleted

You are posting questions as alternative solutions. Please use the 'Have a Question or Comment' link just below the tip to ask questions.

Niklas L 7-Oct-11 12:33pm View

I second that.

Niklas L 6-Oct-11 7:30am View

To OP: In addition, having the information in the model enables you to change your graphic background to something more interesting, like an image or similar, not having to rely on specific color information.

Niklas L 22-Sep-11 7:27am View

Do the Find in Files thing on the wxWindows folder and see how it's all connected.

Niklas L 22-Sep-11 6:41am View

On what line does it fail? Step through the code in a debugger, and see which of your statements that fail.

Niklas L 22-Sep-11 6:26am View

As a note, you will never be able to prevent reverse engeneering, only make it harder. One would still have the possibility of running a debugger on your exe, wait for it to decrypt the secret portion, and then figure out what's in there.

Niklas L 13-Sep-11 9:15am View

This will not reduce dependencies and compile time, which the OP is asking for.

Niklas L 8-Aug-11 18:18pm View

Thanks Albert!

Niklas L 8-Aug-11 7:08am View

Short and precise. 5.

Niklas L 8-Aug-11 4:22am View

There you have it. You can't do a memcpy on those members. Add a copy constructor (and an assignment operator and possibly add move constructor if your compiler allows it). Skip CopyMemory and do a correct copy instance by instance, for example by using std::copy as I suggested.

Niklas L 1-Aug-11 5:45am View

That gave me a few options. Thanks for that!

Niklas L 8-Jul-11 11:59am View

Fived for the example alone.

Niklas L 6-Jul-11 15:22pm View

You can even set a break point in VS to trigger when that variable changes, to quickly find the offending module. 5.

Niklas L 6-Jul-11 9:39am View

Simply because your're an excellent C++ programmer. I imply that you knew before testing it that the inclusion of the member variable wasn't the real problem, and that you also knew your own code would work with or without it. If you could not add a vector<int> without having your code break, not much code would work.
It was meant as a compliment rather than an accusation, even if it was somewhat disguised.

Niklas L 6-Jul-11 7:16am View

Did you actually add it to your own code to verify that? Something tells me you're not being completely honest here :)

Niklas L 1-Jul-11 12:38pm View

Check the specific API on MSDN. The documentation almost always contains info on what lib file to use at the bottom of the page.

Niklas L 1-Jul-11 4:44am View

You cannot render a JPEG directly. It has to be unpacked prior to rendering.

Niklas L 30-Jun-11 6:53am View

I trust the author on the other side of that link :) For every answer you can add to the chain, I‘ll vote 5, starting right now...

Niklas L 21-Jun-11 2:45am View

As a rule of thumb, if you implement the copy constructor, you should implement the assignment operator as well.

Member pointer variables you have allocated memory for, need to be deleted in the destructor. This seems to apply to your implementation.

Niklas L 14-Jun-11 14:28pm View

A good link, although harder to customize than Joan Murt's solution. Fived nevertheless.

Niklas L 14-Jun-11 14:26pm View

'Good enough' is always the best way to solve things. Fived.

Niklas L 13-Jun-11 10:45am View

Have you tried starting it from the Services GUI? Who wrote the service you‘re trying to start? Do you have access to the source? Checking the Event Log is a good idea.

Niklas L 13-Jun-11 8:04am View

Add a ::DebugBreak() and attach the debugger to see what happens.

Niklas L 13-Jun-11 3:54am View

Good explanation. 5.

Niklas L 11-Jun-11 4:18am View

Yes I did read it before, and found it quite entertaining. A very innovative approach indeed.

Niklas L 10-Jun-11 14:22pm View

Thanks. Creating a matrix class would be a minimum.

Niklas L 10-Jun-11 8:19am View

If you know how to calculate the matrix multiplication resulting in a MxN matrix, you can partition the calculation to several (M/y)x(N/y) calculations. Implement that in a separate function before you even try using threads. Then create the threads and make each thread calculate one of the partitions using the function you wrote.

Niklas L 8-Jun-11 9:16am View

There can be no conflict reading data. Are you writing to the array as well? It's impossible to answer your question without more information on how you do it.

Niklas L 8-Jun-11 7:05am View

What makes you think the encoded byte array can be used as a string of characters? E.g. does your Encrypt() method guarantee there are no null bytes in the output? Sounds unlikely.

Niklas L 8-Jun-11 5:54am View

I do understand your point, and technically you're right. I'm saying that the lack language support of this information is a source of errors. If you add a new value to your enumeration sequence, you will also have to update another value somewhere else, the const int Length in your case. You might introduce very hard to catch bugs if you forget this. It's a maintenance issue. This is where constructs like this comes in handy. It's far from a perfect solution, for reasons given, but it works quite well and is common practice in C/C++.

Niklas L 8-Jun-11 5:32am View

"Deduces an error"
What error?

Niklas L 8-Jun-11 3:19am View

This is a good comment and should be used as an incentive to perform a manual system overhaul.

Niklas L 8-Jun-11 3:08am View

You can use the enum values to index arrays, in which case knowing the bounds is crucial. Implementation of a deck of cards is a classical example. It is not likely however, that someone will introduce a new suit with a new version of the software, but imagine the same thing where the enum is representing a state which can be expanded in future versions. Not being able to programatically tell the number of items in the sequence is a great source for errors.

Niklas L 8-Jun-11 2:30am View

IMO it's rather clear that the OP doesn't have BIRTHDATE = 100. The dynamic enums are not dynamic during runtime, but rather during the evolution of the software. I can think of several situations where this comes in handy. Not being able to tell the number of items in an enumerable sequence is a severe limitation.

Niklas L 7-Jun-11 16:48pm View

I agree, thanks.

Niklas L 7-Jun-11 9:16am View

After calling GetWindowRect(), you should call ScreenToClient() to translate from screen coordinates to client coordinates of the parent (i.e. the dialog)

Niklas L 7-Jun-11 8:38am View

Good link :)

Niklas L 7-Jun-11 7:05am View

Good general advice. 5.

Niklas L 7-Jun-11 7:04am View

Fived.

Niklas L 6-Jun-11 15:26pm View

Now that's a bold statement! :) 5!

Niklas L 6-Jun-11 15:22pm View

Someone ate your brackets!! :oops:

Niklas L 1-Jun-11 5:14am View

Thank you!

Niklas L 1-Jun-11 5:14am View

Good link!

Niklas L 1-Jun-11 3:55am View

Actually, I would not recommend that page, since the implementation doesn't check for self-assignment, where the execution would be undefined.

Niklas L 29-May-11 14:41pm View

Thanks.

Niklas L 27-May-11 3:43am View

A well deserved 5.

Niklas L 27-May-11 3:03am View

Thanks.

Niklas L 27-May-11 3:01am View

_T(x) is only a macro which will expand to prefix its argument string with 'L' if UNICODE is defined or just x otherwise. It can be used with or without MFC.

Niklas L 25-May-11 3:51am View

Five for patience :)

Niklas L 25-May-11 3:35am View

Fived, although try/finally is not yet part of C++. A RAII implementation is the closest pattern.

Niklas L 20-May-11 2:52am View

Thanks for that!

Niklas L 20-May-11 2:52am View

Thank you!

Niklas L 20-May-11 2:52am View

Thanks.
Please see my reply to your more complete answer. :)

Niklas L 20-May-11 2:50am View

Maybe a somewhat constructed problem, since you should avoid placing things at file level in the global scope. Hence, you should always be able to qualify your enum values using the enclosing class or namespace.

Niklas L 19-May-11 9:09am View

Good general advice. Fived once more.

Niklas L 19-May-11 9:07am View

Good link indeed! 5.

Niklas L 19-May-11 9:03am View

Sending a five in your general direction.

Niklas L 19-May-11 6:19am View

I was thinking memory mapping and block copying.

Niklas L 19-May-11 3:52am View

Copy and paste the relevant code. And standard headers is not considered relevant code. Follow the stack trace back to your own code. Also, don't post questions as answers, edit your original question instead.

Niklas L 18-May-11 14:46pm View

Traffic control or traffic monitoring?

Niklas L 18-May-11 14:46pm View

...or traffic simulators, road-side detector data collection, floating car data, traffic light adaptive logic, traffic flow analysis or maybe intelligent speed adaptation.

Niklas L 18-May-11 12:13pm View

I was thinking the same, but it's actually 0.66 that will overflow by a fraction. Nevertheless, it's always good to avoid floating point arithmetics in loops, when all you want is to run the loop a specific number of times.

Niklas L 18-May-11 12:05pm View

I added a copy constructor to C, and it still works. Then I added a copy constructor to A, but C can still touch his private parts, so to speak. What did you do to have it not working?

Niklas L 18-May-11 11:55am View

That's a good answer. 5.

Niklas L 18-May-11 11:55am View

Fived.

Niklas L 18-May-11 11:48am View

"if you bring in the address from some control that's usually already handled there", a somewhat strange formulation, since escaping is done only for the compiler to understand it. Once your program run, you don't need to escape backslashes.

Niklas L 18-May-11 11:38am View

It does make life easier. Fived.

Niklas L 18-May-11 8:24am View

Your question makes no sense. Unique compared to what? Start with 1000000000 and add 1 each time. That will give you all possible ten digit unique numbers. (A 32 bit representation is not enough, so pick a 64 bit integer.)

Niklas L 18-May-11 4:08am View

But of course. Unless you are running commands from a linux prompt. ;)

Niklas L 18-May-11 3:51am View

I was referring to the OP taking a course or not. And yes, it would be stupid, but only if you knew there was something called a debugger.

Niklas L 18-May-11 3:09am View

Thanks. Sometimes I get the impression that they don't cover debugging techniques at programming courses at all. Some might learn programming on their own, but there are clearly a lot of students asking basic questions here, which could have been answered by a debugger. I'm not saying this one is such though.

Niklas L 18-May-11 3:08am View

Deleted

Thanks. Sometimes I get the impression that they don't cover debugging techniques at programming courses at all. Some might learn programming on their own, but there are clearly a lot of students asking basic questions here, which could have been answered by a debugger. I'm not saying this one is such though..

Niklas L 16-May-11 13:40pm View

Merci!

Niklas L 16-May-11 7:34am View

Fived.

Niklas L 16-May-11 6:47am View

Fived!

Niklas L 16-May-11 6:46am View

Fived. Again.

Niklas L 13-May-11 3:28am View

What are you talking about? What symbols?
1: Where is your image?
2. What exactly is it that makes it hard to save it to the database?

Maybe you should update your question with a small code snippet, showing what you have done?

Niklas L 11-May-11 7:26am View

Don't you have a jpeg image? Is it located on disk? If so read the entire file to a chunk of memory. There you have it. The BLOB.

Niklas L 11-May-11 6:46am View

Don't you have a pointer to the jpeg image? That will be the BLOB.

Niklas L 11-May-11 6:43am View

United Kingdom: 5 points. Royaume-Uni: cinq points.

Niklas L 11-May-11 6:39am View

If you want to parse files larger than 2Gbyte, you will need a different approach. You should not read it all at once, but rather just read the portion you are currently working on, process it, and read a new portion.

Niklas L 11-May-11 6:30am View

You got a five here, even though you'll have to fix your first code block. It's a bit XML-ish. :)

Niklas L 10-May-11 14:45pm View

Ah! So you also had that assignment! :p

Niklas L 10-May-11 9:28am View

Don't add comments as solutions. Either click on the 'Add comment' link below an answer if you want to reply to it, or hit 'Improve question' if you have more info you want to share.

Niklas L 10-May-11 9:23am View

Comment from OP:
"no i didn't..
it should be done till friday..
if someone know please write"

"i'll try something,
when i stuck somewhere i'll post here"

Niklas L 10-May-11 0:51am View

Fived.

Niklas L 9-May-11 3:34am View

Changed tags.

Niklas L 6-May-11 4:31am View

Very good comment. 5.

Niklas L 6-May-11 4:29am View

Fived.

Niklas L 6-May-11 4:08am View

Nothing wrong with this answer, and it's perfectly valid. Fixing the score.

Niklas L 6-May-11 2:46am View

This is a very likely cause for trouble.

Niklas L 6-May-11 2:34am View

Is this MFC or plain Win32?

Niklas L 6-May-11 2:33am View

Thanks

Niklas L 4-May-11 8:37am View

That's very kind of you Panili. ;)

Niklas L 4-May-11 8:34am View

Thanks

Niklas L 4-May-11 7:19am View

What makes you think you can index an uninitialized variable? (projects in while-loop)

Niklas L 4-May-11 7:10am View

I guess the easiest possible solution. Fived.

Niklas L 2-May-11 15:46pm View

This should have been marked as an answer by OP. My 5.

Niklas L 2-May-11 15:40pm View

Good call! 5.

Niklas L 2-May-11 14:24pm View

Backwards is the key to success :) Fived.

Niklas L 2-May-11 14:20pm View

Good recommendation. 5. If OP wants to change the values using a hook, it's probably nothing I want to have installed on my machine.

Niklas L 2-May-11 14:17pm View

Fived.

Niklas L 2-May-11 11:16am View

Not really necessary to give away how to get rid of const-ness, it's there for a reason, but 5 nonetheless.

Niklas L 2-May-11 11:08am View

Fair enough. It smells homework.

Niklas L 2-May-11 11:02am View

Added to the fives.

Niklas L 2-May-11 4:06am View

You're welcome

Niklas L 2-May-11 2:44am View

You're welcome.
fopen_s(&MenuFile,"D:\\mlkcfg.txt", "wb");
would also work.

Niklas L 2-May-11 2:27am View

I fail to see what a (correct) statement about C++ has to do with any other language.

Niklas L 1-May-11 11:20am View

Quite right in theory. Most notably cross module optimizations and use of target instruction set. I don't know how much of this is actually implemented in the JIT compiler though.

Niklas L 30-Apr-11 4:25am View

Just out of curiosity: Why do you read and write binary data to a file opened in text mode?

Niklas L 29-Apr-11 7:18am View

Had a look, and adding something like if (boost::is_pod<T>()) {} else {}</t> will only add to compile time, and will leave no traces in the optimized object code. The optimizer will remove the entire branch that isn't used. So it might just be worth it.
Edit: Template argument was eaten by the html parser.

Niklas L 29-Apr-11 6:21am View

Thanks. My psychic powers are not what they used to be :)

Niklas L 29-Apr-11 6:17am View

Ah, ok. Did you check the boost implementation of is_pod and is_pod_array?
http://www.boost.org/doc/libs/1_43_0/boost/type_traits/is_pod.hpp

Niklas L 29-Apr-11 5:28am View

Both constructors and destructors of POD types are no-ops. And you can use constructor semantics to initialize POD types as well. int a(0);

Niklas L 29-Apr-11 4:29am View

Or better yet:
const wstring& getName() const {...}

Niklas L 29-Apr-11 3:49am View

This is the best solution. Don't give the user a chance to make a mistake. High 5.

Niklas L 29-Apr-11 3:45am View

Good link to a simple solution. Consider yourself fived.

Niklas L 28-Apr-11 8:01am View

Thanks! I found a boost implementation:
http://www.boost.org/doc/libs/1_38_0/boost/rational.hpp

Niklas L 28-Apr-11 4:08am View

If you can compile the service, and then start it successfully, it's not an MFC related problem.

Niklas L 28-Apr-11 4:07am View

How do you get hold of your hScrDC? If I were you I would try running the service as the interactive user. It might be a security issue.

Niklas L 28-Apr-11 2:45am View

"alternatively educated" :D

Niklas L 27-Apr-11 14:08pm View

Don't worry. I was thinking the same thing when I saw it :)

Niklas L 26-Apr-11 16:57pm View

There was no break in the original version. The question was updated, and the Griff was right.

Niklas L 26-Apr-11 8:03am View

Good advice. 5.

Niklas L 26-Apr-11 4:26am View

Good links, but the first two does not apply to VS VC 5.0. None the less a 5.

Niklas L 26-Apr-11 4:11am View

Very likely cause. 5.

Niklas L 26-Apr-11 4:08am View

5 for your remote debugging abilities :)

Niklas L 21-Apr-11 6:40am View

Your own font trick worked remarkably well ;p

Niklas L 21-Apr-11 6:26am View

Then the index won't be higher than 98.
<font='tiny and white'>I changed the loop conditions before writing my comment above</font>

Niklas L 21-Apr-11 5:35am View

I don't know what you are talking about :rolleyes:

Niklas L 21-Apr-11 3:18am View

You give the path of something that looks like a file stream, but the error message is about a string stream. Sounds confusing. Are there no string stream headers you can use?

Niklas L 21-Apr-11 3:01am View

Deleted

Most code can be rewritten to be faster. But almost always, this is a trade-off, sacrificing readability and maintainability. Unless you have real-time requirements, gaining those few percents just isn't worth it. Your code is clean and easy to follow, which is why I can spot a potential memory leak :) See? Maintainability again! You need to delete pBitmapBits just before your second return statement. (You don't have to check for null before deleting, since delete will do that for you.)

Niklas L 20-Apr-11 15:15pm View

There is a difference. The lpBits. He allocates bm.bmHeight * bm.bmWidth bytes, when he should allocate an amount equal to bm.bmHeight * bm.bmWidthBytes.

And now I see another problem with the code. He is writing to the buffer after he creates the bitmap.

Niklas L 20-Apr-11 14:45pm View

So what do you use to back that up? There are plenty more statements on MSDN claiming 16 bit alignment for up to 16 bits per pixel. Hera are a few: http://msdn.microsoft.com/en-us/library/k1sf4cx2.aspx

Also, you state that he cannot use a modified width. In fact he has to, and since you claim DWORD alignment, even more so. What if his original bitmap is 3 pixels wide? He can't use 3 bytes per scan line, as will be the case if he doesn't modify the width. Have a look at how he allocates the pixel buffer.

Notice that he uses two separate bitmaps.

Niklas L 20-Apr-11 13:44pm View

From MSDN on CreateBitmap():
"lpvBits [in]

A pointer to an array of color data used to set the colors in a rectangle of pixels. Each scan line in the rectangle must be word aligned (scan lines that are not word aligned must be padded with zeros). If this parameter is NULL, the contents of the new bitmap is undefined.
"

Niklas L 20-Apr-11 7:34am View

Deleted

I would not worry about people claiming there will be cache misses, or that you should use multiplication or shift operations instead of division. They either have no insight in optimization techniques, or they haven't upgraded their compilers since the last millennium. Some myths die hard. Consider yourself fived.

Niklas L 20-Apr-11 7:07am View

Deleted

I second thatraja. And I also think the Font should be a parameter instead of fontname ant fontsize to allow for font decoration and such.

Niklas L 20-Apr-11 6:53am View

Nice catch!

Niklas L 20-Apr-11 6:41am View

Every programmer should read that book. Fived.

Niklas L 20-Apr-11 4:17am View

There might just be something to it then :)

Niklas L 20-Apr-11 4:05am View

Sounds reasonable. 5 for psychic power.

Niklas L 19-Apr-11 7:39am View

I think he's trying to rid himself of MFC as per his other questions.

Niklas L 19-Apr-11 4:54am View

Hehe, yes, Hans should be payed by Mr Brooks :)

Niklas L 19-Apr-11 4:17am View

That makes me want to read the book! Not pay for it, but read it, and I will have your comment in mind while doing it. I will put it on my todo list.

Niklas L 19-Apr-11 4:08am View

Do you want to use MFC to build a windows service? MFC is mainly UI classes, which doesn't go with services very well. Could you give some more input?

Niklas L 19-Apr-11 3:01am View

Well, it is MFC after all. Actually, the worst way is to use a static char*. The example on MSDN uses a static char*, and it's easy to use. But what happens the day you want to localize your app? A CString::LoadString() would be a lot better. But I agree, I prefer std::string over CString whenever I can. Thanks for the 5.

Niklas L 18-Apr-11 15:36pm View

Indeed it is

Niklas L 18-Apr-11 14:21pm View

If you right-click it and pick 'Go To Declaration'. Can't VS find the declaration then? (It does miss on occasion.)

Niklas L 18-Apr-11 14:06pm View

Good advice. Compensating the down votes a tad.

Niklas L 18-Apr-11 12:06pm View

Yes, but with strncpy() being the odd exception.

Nice link btw.

Niklas L 18-Apr-11 11:33am View

Ah! Lost in translation. Sorry. ...and thanks!

Niklas L 18-Apr-11 11:24am View

Good answer! 5!

Niklas L 18-Apr-11 11:21am View

And why is it not a good idea to change the function if you get a warning? The warnings are there for a reason. Too many programmers have made too many errors involving strcpy and buffer overflows over the years. Microsoft decided to do something about it, promoting the new 'secure' versions of the old string manipulation functions. This is actually what Microsoft recommends.

Niklas L 18-Apr-11 11:08am View

strncpy() does not guarantee a null terminator in the target string if the buffer is full which is a real problem. So strncpy() is also tagged insecure.
The template versions of strcpy_s() working on arrays are quite nice however.

Niklas L 18-Apr-11 2:05am View

Thank you!

Niklas L 18-Apr-11 2:05am View

Thanks!

Niklas L 18-Apr-11 2:03am View

Hehe, sorry. Use wcout instead of cout.

Niklas L 15-Apr-11 14:48pm View

IterAvi->c_str()

Niklas L 15-Apr-11 9:29am View

Although a working solution, you will pollute your app class with something that isn‘t it‘s concern, adding to build dependencies and compile time.

Niklas L 14-Apr-11 15:17pm View

Unless you want to write your own filebuf implementation, I'd say yes.

Niklas L 14-Apr-11 6:51am View

That is a strange thing to ask at an interview. At a geek get-together yes, but at an interview... hmm...

Niklas L 11-Apr-11 14:28pm View

API or keyword, there's only a time-to-market difference. You could one day decide that the use of for-loops should be avoided, in favor of while-loops. Every new compiler would issue a warning on for-loop use. Not as easy as pushing new API sources I admit, but not unrealistic.
You might argue that these changes never comes to those who doesn't upgrade to the latest compiler, but that also holds for those not upgrading to latest API version. There really isn't a big difference, other than time-to-market.
(This comment is for API vs keyword only.)

Niklas L 11-Apr-11 6:33am View

Check out FindFirstFile() and FindNextFile() API on MSDN.

Niklas L 11-Apr-11 5:31am View

You can make a keyword deprecated and still allow using it for backward compatibility.
But I wish you all the best in your quest for a better 'override' term :)

Niklas L 11-Apr-11 5:01am View

It would probably be easier to change the name if it did appear in code. Its hard to tag a commonly used term deprecated.

Niklas L 11-Apr-11 4:50am View

I see that as a convenience for the programmer, not a necessity for OOP as a concept.

Niklas L 11-Apr-11 4:43am View

As with all aspects of programming, good naming is essential for learning, readability, maintainability, etc. Unfortunately, this one cannot easily be re-factored.

Niklas L 11-Apr-11 4:31am View

Good answer! Overloading and overriding is too often confused seeing the questions asked here.

Niklas L 7-Apr-11 3:21am View

Inheritance would imply the *this instance.

Niklas L 7-Apr-11 3:15am View

If you send a nWindowSize of zero to your constructor, your code will break. Why do you send nPointSize to the constructor, and then again to Update and overwite the old value? Could you not remove it from the constructor, or do you have more code not shown here?

Niklas L 4-Apr-11 4:10am View

Ok, well then that won't be a problem.

(I just read "the content of stream is a simple 16 bits counter" meant sending 16 bits in each packet.)

Niklas L 4-Apr-11 3:16am View

The minimum UDP packet size for sending a 16 bit value is 80 bits, of which the header is 64 bits. That's quite a lot of overhead.

If the force is with you, you could maybe use the header checksum field to store your data, but that will still require 64 bits to send a 16 bit value.

Niklas L 31-Mar-11 3:15am View

5 for your effort, and 5 for MNMR's attitude towards learning

Niklas L 31-Mar-11 2:59am View

It's from homework, or a job interview, and yet another repost of http://www.codeproject.com/Questions/174563/Basic-of-C-and-Cplusplus-quiz-Questions.aspx

Niklas L 31-Mar-11 2:54am View

If one more of these comes along, I will start to wonder if I missed the latest software development trend, and search for a book on amazon. :)

Niklas L 30-Mar-11 16:20pm View

If I were you, I would be more interested in examining the last function in your project. Also, you should check if there are any relevant differences in pre-processor directives in debug vs release.

Niklas L 30-Mar-11 7:00am View

It only points to the first character. It doesn't point to any other characters in the sequence.

Niklas L 29-Mar-11 3:11am View

Adjusted a bit, babelfish needs, no? :)

Niklas L 29-Mar-11 2:34am View

You do have access to the CListCtrl somewhere. If it is a CListView, use CListView::GetListCtrl() as hansen described, and if it is a CListCtrl in a dialog of CFormView, just add a variable for it, or access it with GetDlgItem(IDC_MYLIST_WHATEVER).

Comments by Niklas L (Top 200 by date)