Re: again for InterlockedIncrement() - C / C++ / MFC Discussion Boards

Matthew Faithfull3-Jun-07 1:18

3-Jun-07 1:18

Are you looking for a compiler? Or wanting to know how to change your build settings to get an EXE as opposed to a DLL? Unsure | :~

Nothing is exactly what it seems but everything with seems can be unpicked.

Re: DLL and EXE files

prithaa3-Jun-07 1:42

prithaa

3-Jun-07 1:42

Hello,
Yes I have made classes through VC++ MFC AppWizard(exe)option in non Doc-View architecture.
Now I want these classes to be used in my another project without actually disturbing the classes.For this purpose if I can make the .dlls and load them in my project directory and #include these .h in my project I think that will be fine.
For now what I do is Add those files in my project through the workspace area and I can actually see all the files in my workspace which I don't want.

Can do like that or I have missed something while making the classes?
Thanks for your replies
Pritaa

Re: DLL and EXE files

Matthew Faithfull3-Jun-07 2:59

Matthew Faithfull

3-Jun-07 2:59

If you've made the classes through the exe AppWizard then it isn't going to be straight forward to try and package them in a Dll. I would suggest starting a new project, An empty Dll that supports/(is linked to) MFC. Then copy the classes you want into it starting with the simplest ones. Then you need to look into exporting the classes from the dll that you want to use. Check this page for a start.
http://msdn2.microsoft.com/en-us/library/3y1sfaz2(VS.71).aspx[^]
I would recommend experimenting a little with creating simple Dlls and using them from simple EXEs before bringing MFC into the picture. Because MFC tries to be a framework it has to assume a 'model' ie EXE or DLL when is generates application classes and the code you get is targetted at one or the other. Things like individual Dialog classes are easy enough to move into a Dll but the main Doc/View classes won't be. Good luck with it anyway Smile | :)

Nothing is exactly what it seems but everything with seems can be unpicked.

Re: DLL and EXE files

Hamid_RT3-Jun-07 7:33

Hamid_RT

3-Jun-07 7:33

Do you need to Creating and consuming MFC DLLs for Beginners[^]

WhiteSky

How do I make a child dialog draw on a different HDC

KellyR2-Jun-07 17:51

KellyR

2-Jun-07 17:51

Hi,

So here's my problem:

I have a child dialog on my main window. I use double-buffering in my application and I'd like the dialog to draw on my back-buffered HDC instead of my front HDC (the HDC of the main window) to prevent flicker when it invalidates.

How can I change the HDC that the child window draws to?

Thanks!

Re: How do I make a child dialog draw on a different HDC

Naveen3-Jun-07 16:17

Naveen

3-Jun-07 16:17

KellyR wrote:
How can I change the HDC that the child window draws to?

Try specifiing the CS_OWNDC style to the child dialog.( using the SetClassLong() function )

nave

[OpenedFileFinder]

again for InterlockedIncrement()

includeh102-Jun-07 16:52

includeh10

2-Jun-07 16:52

long iv=0;
function InterlockedIncrement(&iv) is used for locking iv according its name, but variable iv is not locked after the function call.

could u explain in which case the iv is locked by the function call?

Re: again for InterlockedIncrement()

Matthew Faithfull3-Jun-07 1:07

Matthew Faithfull

3-Jun-07 1:07

iy = InterlockedIncrement(&iv); is essentially the same thing as

<br />
SomeCriticalSection.Lock();<br />
iy = (++iv);<br />
SomeCriticalSection.Unlock();<br />

The reason you would use InterlockedIncrement is if iv is being accessed by more than one thread, as you know, and you want to increment and get the value of iv with no possibility of another thread steping in and changing it in the mean time. Consider one possible break down of the unprotected code

iy = iv++;

in 'optimised' pseudo assembler Big Grin | :-D

1 load iv into a register
2 If iv is an even number
3 set bit 1 on the register
4 else
5 do a full ADD operation on 1 and iv
6 store the register contents back again into the memory of iv
//..one critical weak point is here
7 load iv from memory again
8 store it a second time into the memory of iy

In theory at least another thread could jump in at the critical weak point in the lines above, and do any amount of stuff with iv in memory D'Oh! | :doh:

Total disaster can then ensue. In practice this is unlikely because a decent compiler backend will optimise out the storing and reloading of iv before assignment to iy. It will just store the same value twice.
The reason you would still use the InterlockedIncrement is not to protect this thread but to protect the other thread which while we are stopped on step 5 writes iv (does it's own step 6) and then finds that it hasn't been written because we jump back in and write over it in our step 6. The iv and iy in the other thread will then be out of sync. Smile | :)

I have found very few situations where InterlockedIncrement is really needed but there are a few.

Nothing is exactly what it seems but everything with seems can be unpicked.

Re: again for InterlockedIncrement()

Mark Salsbery4-Jun-07 5:40

Mark Salsbery

4-Jun-07 5:40

I explained this to the OP in another post.

While you're correct about locking (probably) not needed on a simple increment,
InterlockedIncrement does more:

SomeCriticalSection.Lock();
iy = (++iv);
returnval = iv;
SomeCriticalSection.Unlock();
return returnval;

Aother thread incrementing iv after the first thread increments it but before the first thread
has copied it could have an undesired effect Smile | :)

Especially on reference counters, which is (probably) most often the reason InterlockedIncrement
and InterlockedDecrement are used.

Mark

"Posting a VB.NET question in the C++ forum will end in tears." Chris Maunder

Re: again for InterlockedIncrement()

Mark Salsbery4-Jun-07 5:45

Mark Salsbery

4-Jun-07 5:45

The variable is locked DURING the call, not after the call completes.

See here[^].

As Matthew mentioned, what's going on in the call is something like:

LONG InterlockedIncrement(LONG volatile* lpAddend)
{
   SomeCriticalSection.Lock();
   ++(*lpAddend);
   LONG returnval = *lpAddend;
   SomeCriticalSection.Unlock();
   return returnval;
}

Mark

"Posting a VB.NET question in the C++ forum will end in tears." Chris Maunder

Suggestion: Subject Matter Editors

Jeffrey Walton2-Jun-07 11:46

Jeffrey Walton

2-Jun-07 11:46

Hi All,

I know this has come up in the context of non-English speaking contributors in an effort to improve article quality. I'm all for it. I submit that the non-English speaking author has a better command of English than I have of his native tongue. So I am humbled by my own ignorance...

However, Subject Matter is a different story. I have a good understanding in some arenas. This knowledge extends sufficiently so that I feel I can provide feedbck and advice. One such area is Cryptography.

I am appaled at some of the articles submitted on Cryptography. In the following, I am not singling out the authors. I am trying to make a succint point.

The latest is TrueCrypt Explorer[^]. This library is using source code and presumably home grown ciphers from the notorius Tom St. Dennis of sci.crypt. Single handedly, he nearly ruined the usenet group with his arrogance and ignorance.

I assume the author is not versed on Cryptography, and as such has fallen victim to TSD's 'Proof by Intimidation' of his cryptosystem. I am not aware of a Cryptanalytic effort by any reputable cryptographer on his work.

Build your own cryptographically safe server/client protocol[^] is misleading because the author states something to the nature he is collecting 'true randomness', which is generally not possible in software.

CTrueRandom - Getting True Random Numbers[^] claims it generates 'True Random Numbers'. I assume the author meant nondeterminsitic method, similar to the noise resisitor placed on the hardware based PRNG board of a RS6000.

As a final example, I point you toward The Art & Science of Storing Passwords[^], where the article claims the practice of storing a password for recover on disk unencrypted is OK.

Jeff

Re: Suggestion: Subject Matter Editors

Hans Dietrich3-Jun-07 0:14

Hans Dietrich

3-Jun-07 0:14

This is an interesting idea, and I too have felt really aggravated by some of the poor articles posted here. I'm sure you know that historically, the CP editors take care of compliance with formatting, etc., but rarely get involved with content, unless there is a serious concern about plagiarism or something similar. I think that over time, the better articles are recognized and get better ratings.

But SMEs are a whole new thing. I can see how putting content responsibility in the hands of a few anointed SMEs could lead to personality problems just like the one you described. A heavy-handed SME could cause huge damage to the site's collaborative spirit (I'm thinking of another Usenet newsgroup, which is dominated by an obnoxious person who was beaten off when he tried the same stuff here). I think I would rather risk a few bad articles than having to contend with a situation like that.

Why don't you think that submitting (positive) comments on the article forums is the way to go?

Best wishes,
Hans

[CodeProject Forum Guidelines] [How To Ask A Question] [My Articles]

Re: Suggestion: Subject Matter Editors

Matthew Faithfull3-Jun-07 1:12

Matthew Faithfull

3-Jun-07 1:12

Could the voting system be tweeked so that an article can be voted into oblivion by, for example, 'nill point' votes from 3 separate experts/moderators/SME? This might stop or at least slow down one bad egg from causing trouble but would allow a way to can really bad articles.

Nothing is exactly what it seems but everything with seems can be unpicked.

Re: Suggestion: Subject Matter Editors

Cliff Hatch3-Jun-07 2:32

Cliff Hatch

3-Jun-07 2:32

Hans

Apologies, I intended to vote your comment 5, but somehow managed to click 1 by mistake Blush | :O

- do you know any way to retract a vote?

I didn't particularly want to participate in this debate, but I agree that the submission of positive comments is a good strategy.

Best Regards

Cliff

Re: Suggestion: Subject Matter Editors

Hans Dietrich3-Jun-07 4:14

Hans Dietrich

3-Jun-07 4:14

Cliff Hatch wrote:
do you know any way to retract a vote?

I think you have to send an email to sean@codeproject.com.

Thanks for telling me - now I won't think that Jeff is being mean. Wink | ;)

Best wishes,
Hans

[CodeProject Forum Guidelines] [How To Ask A Question] [My Articles]

Re: Suggestion: Subject Matter Editors

Jeffrey Walton3-Jun-07 9:35

Jeffrey Walton

3-Jun-07 9:35

Cliff Hatch wrote:
Apologies, I intended to vote your comment 5, but somehow managed to click 1 by mistake - do you know any way to retract a vote?

Perhaps Chris can help. I would post a questiion on the suggestion board (where this suggestion should have gone - I don't know how I managed to insert it into C++).

Jeff

Re: Suggestion: Subject Matter Editors

Jeffrey Walton3-Jun-07 9:33

Jeffrey Walton

3-Jun-07 9:33

Hi Hans,

Hans Dietrich wrote:
I can see how putting content responsibility in the hands of a few anointed SMEs could lead to personality problems just like the one you described. A heavy-handed SME could cause huge damage to the site's collaborative spirit (I'm thinking of another Usenet newsgroup, which is dominated by an obnoxious person who was beaten off when he tried the same stuff here). I think I would rather risk a few bad articles than having to contend with a situation like that.

As Gary raised the same issue below, I was not not trying to create censorship.

I was thinking more the way a Book Proofer work: Make suggestions and comment on tecnical correctness. The author can choose to incorporate.

Jeff

Re: Suggestion: Subject Matter Editors

Gary R. Wheeler3-Jun-07 1:52

Gary R. Wheeler

3-Jun-07 1:52

The CodeProject is an informal, open, collaborative site. The approach you describe would convert it to a formal, closed, scholastic one. Demi-gods would 'vet' articles based on their own notion of accuracy, usefulness, and appropriateness.

I think the current approach with the article voting system and each article having its own forum is sufficient.

If you think these articles are poor articles, vote them down. Post a message in their forum stating THIS ARTICLE IS CRAP and explain why. Write a response article.

Software Zen: delete this;

Fold With Us![^]

Re: Suggestion: Subject Matter Editors

Jeffrey Walton3-Jun-07 9:28

Jeffrey Walton

3-Jun-07 9:28

Hi Gary,

Gary R. Wheeler wrote:
The CodeProject is an informal, open, collaborative site. The approach you describe would convert it to a formal, closed, scholastic one. Demi-gods would 'vet' articles based on their own notion of accuracy, usefulness, and appropriateness.

The situation you describe (censorship) is worst than the current. I was not imagining an involuntary system.

Perhaps middle ground is the solution: the Content Editor makes suggestions. If the author chooses, he or she can incorporate the suggestions.

Gary R. Wheeler wrote:
If you think these articles are poor articles, vote them down. Post a message in their forum stating THIS ARTICLE IS CRAP and explain why.

I try not to do this (I think it is very unprofessional). Again, I'd rather correct the problem at the source...

Gary R. Wheeler wrote:
Write a response article.

In my case, I did: A Survey of Pseudo Random Number Generators[^]. Also, I did try to solicit feedback from two PhD's. Neither would comment. In either case, I tried by going to someone who knows more then me.

Humorously, the beginner article has only garnered a 4/5 rating

Jeff

NTFS Alternate Data Streams

Akin Ocal2-Jun-07 10:26

Akin Ocal

2-Jun-07 10:26

I want to get buffers of all streams of a file. I searched msdn then i found ready functions

FindFirstStream and FindNextStream which i ve found in Windows Platform SDK.

These functions give me name of streams. So how can get data in these streams using their name ?

I guess i will use CreateFile and ReadFile again but , what must the parameter given to CreateFile , must it be only stream name or "filenametreamname" ??

Thank you very much...

Re: NTFS Alternate Data Streams

Mark Salsbery2-Jun-07 15:17

Mark Salsbery

2-Jun-07 15:17

Akin Ocal wrote:
These functions give me name of streams. So how can get data in these streams using their name ?

You can append a colon and the stream name to the pathname of the file.

filenamename.ext:streamname

Mark

"Posting a VB.NET question in the C++ forum will end in tears." Chris Maunder

log10 ( 3/2 ) = 0.000 Help me! :( [modified] Solved

Immunity182-Jun-07 8:14

Immunity18

2-Jun-07 8:14

Hello in my code i have this

	double idf[50];
	for (int zz = 0  ; zz < m_QueryWords ; zz++ )
	{

		idf[zz] = log10 ( (double)(nDoc/df[zz]) );
->	}

While i debug i have :

<code>name		Value
df[zz]		2
idf[zz]		0.00000   (red colour just took the value)
nDoc		3
+this		etc :P
zz		0 </code>

what i do wrong ?

I changed the code with

int location = f + (m_QueryWords*ii);
SCs = SCs + nArray.GetAt(location)*idf[f]*idf[f] ;

and now works
-- modified at 14:32 Saturday 2nd June, 2007

Re: log10 ( 3/2 ) = 0.000 Help me! :( [modified] Solved

Christian Graus2-Jun-07 21:55

Christian Graus

2-Jun-07 21:55

Immunity18 wrote:
idf[zz] = log10 ( (double)(nDoc/df[zz]) );

Change to log10 ( (nDoc/(double)df[zz]) );

df[zz] needs to be a double in order to provide a floating point result.

Christian Graus - Microsoft MVP - C++
Metal Musings - Rex and my new metal blog
"I am working on a project that will convert a FORTRAN code to corresponding C++ code.I am not aware of FORTRAN syntax" ( spotted in the C++/CLI forum )

managed/unmanaged code

includeh102-Jun-07 6:09

includeh10

2-Jun-07 6:09

I posted a question before and got hint of "managed/unmanaged code".

I do heard "Managed C++", I guess it is another language - at least it is different from c++ hear.

my question is:
how to generate "managed code" in our C++, such as in MFC or Win32?
any samples if it is possible?

thx

Re: managed/unmanaged code

Sathesh Sakthivel2-Jun-07 7:39

Sathesh Sakthivel

2-Jun-07 7:39

Hope this article will help you.

Managed Code[^]

Regards,

Satips. Rose | [Rose]

Use Ctrl+Left/Right to switch messages, Ctrl+Up/Down to switch threads, Ctrl+Shift+Left/Right to switch pages.