C / C++ / MFC

Directory Contents Question

18-Dec-03 17:47

Look at SetItemState().

INTP

Roger Wright18-Dec-03 16:17

18-Dec-03 16:17

I've spent the past two days playing with sample code that is supposed to, given a main directory, allow me to select one or more subdirectories and collect stats on them including the number of files, and the total size of each subdirectory. The technique used starts with

hFind = FindFirstFile(strFind, &dataFind);

where strFind is a wildcard search string on the main directory, and dataFind is a WIN32_FIND_DATA structure. Directories are identified by

if(dataFind.dwFileAttributes == FILE_ATTRIBUTE_DIRECTORY)

in a loop that traverses the entire directory. I found that this correctly identified only a few subdirectories, but couldn't figure out why. In a later section, files within a subdirectory are found using the same technique but using

if(dataFind.dwFileAttributes == FILE_ATTRIBUTE_ARCHIVE)

as the search criterion. This, too, failed to find all the files in the subdirectory. After many hours of research and experimenting I deduced that this is a stupid way to do this task, though it was a great way to demonstrate how to use the various controls in a dialog box. The reason, it appears to me, is that the field dwFileAttributes is an OR of one or more of many attributes, and rarely just a single one, so any single value comparison on a multiple-attribute entity will fail.

So my question is, what is a proper method for reliably testing whether a directory entity is a file or a subdirectory?

"Another day done - All targets met; all systems fully operational; all customers satisfied; all staff keen and well motivated; all pigs fed and ready to fly" - Jennie A.

Rob Manderson18-Dec-03 16:55

Rob Manderson

18-Dec-03 16:55

You do it like this

if (dataFind.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY)

and

if (dataFind.dwFileAttributes & FILE_ATTRIBUTE_ARCHIVE)

in other words, and out the bit you're interested in. I wouldn't rely on the archive bit though, it'll miss lots of files. Use FILE_ATTRIBUTE_NORMAL instead.

Rob Manderson

http://www.mindprobes.net

Paul Watson wrote:What sense would you most dislike loosing?

Ian Darling replied.

Telepathy

Then I'd no longer be able to find out everyones dirty little secrets The Lounge, December 4 2003

Roger Wright18-Dec-03 20:38

18-Dec-03 20:38

Thanks, Rob... That was sorta where my small mind was headed - performing an AND operation to isolate the particular bit of interest, then testing for equality. It's unfortunate that the authors of this book didn't take the opportunity in this example to demonstrate this option. I think it would have been very instructive to a lot of people trying to learn the language. At the very least, it would have alerted students to the fact that the Win32 and MFC APIs are full of such flag fields, and that a programmer must learn appropriate methods for dealing with them. To their credit, though, I must admit that the intent was to demonstrate how to populate the fields of several control objects, which the chapter did quite well. Maybe I'm too much a perfectionist, but I think that sample code in a textbook should be perfect. Unfortunately it is often a far cry from merely good, let alone perfect.

"Another day done - All targets met; all systems fully operational; all customers satisfied; all staff keen and well motivated; all pigs fed and ready to fly" - Jennie A.

Rob Manderson18-Dec-03 22:22

Rob Manderson

18-Dec-03 22:22

You're welcome Smile | :)

Which book are you learning from?

I'm learning C# right now - with Tom Archers Inside C# 2nd edition as my guide. Great textbook so far but I didn't fully understand his description of assemblies, manifests et al in chapter 1. No matter - I'll go back and reread that part again and again, on the assumption that I've missed something, when I need to know it.

No textbook will be perfect - the author will always make some assumption somewhere that something is so obvious that it needs no explanation. In this case I suspect it was that the constants are named FILE_ATTRIBUTE_??? and the author assumed no one would imagine they were discrete constants.

Rob Manderson

http://www.mindprobes.net

Paul Watson wrote:What sense would you most dislike loosing?

Ian Darling replied.

Telepathy

Then I'd no longer be able to find out everyones dirty little secrets The Lounge, December 4 2003

Roger Wright19-Dec-03 5:23

19-Dec-03 5:23

Rob Manderson wrote:
Which book are you learning from?

Practical Visual C++ 6, by Bates and Tompkins. It's quite good, actually, and as I mentioned, the intent of the program was to illustrate how to fill up various controls, not to enumerate all the files in a directory. I was just thinking ahead a bit. I'm looking forward to reading Tom's book, but there's no rush as no one around here is using C# yet.

"Another day done - All targets met; all systems fully operational; all customers satisfied; all staff keen and well motivated; all pigs fed and ready to fly" - Jennie A.

John R. Shaw18-Dec-03 17:09

18-Dec-03 17:09

Your conclusion is correct, that dwFileAttributes is a set of bit flags.
This is just of the top of my head.

if( !(dataFind.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY) )
{
    // should be a file
}
else
{
   // should be a file}
}

// TESING MULTIPLE ATTRIBUTES
if( dataFind.dwFileAttributes & (FILE_ATTRIBUTE_ARCHIVE | FILE_ATTRIBUTE_NORMAL) )
{
    // this is a normal or archived file
}

INTP

Roger Wright18-Dec-03 20:45

18-Dec-03 20:45

John R. Shaw wrote:
if( !(dataFind.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY) )
{ // should be a file}
else
{ // should be a file}}

Off the top of your head that's quite reasonable - I understand the intent, despite the typo. Rob came up with much the same conclusion (see above). Thank you for your guidance - C++ is fairly new to me, but I'm finding that it's not really as bad as I'd expected. Big Grin | :-D

"Another day done - All targets met; all systems fully operational; all customers satisfied; all staff keen and well motivated; all pigs fed and ready to fly" - Jennie A.

John R. Shaw18-Dec-03 22:12

18-Dec-03 22:12

You are welcome! I was wondering where that curley brace went to, but it is not a typo since it ended up as part of the comment. By the way, C++ is derived from C and what you are seeing is actualy C code (since no classes are invalved).

Have a great night, day, or whaterver!

INTP

Rob Manderson18-Dec-03 22:55

Rob Manderson

18-Dec-03 22:55

y'know I had to stare at the code John posted quite a long time before I saw the typo Smile | :)

Here I was scratching my head and wondering what Roger was on (and if I could have some Smile | :)

) before finally seeing that errant brace.

Rob Manderson

http://www.mindprobes.net

Paul Watson wrote:What sense would you most dislike loosing?

Ian Darling replied.

Telepathy

Then I'd no longer be able to find out everyones dirty little secrets The Lounge, December 4 2003

Roger Wright19-Dec-03 5:16

GDI Objects LEAK due to manifest file codes?

19-Dec-03 5:16

It was a little more than the curly brace:

if( !(dataFind.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY) )
{ // should be a ~~file~~ directory}
else
{ // should be a file}}

Smile | :)

"Another day done - All targets met; all systems fully operational; all customers satisfied; all staff keen and well motivated; all pigs fed and ready to fly" - Jennie A.

Yu Zhiquan18-Dec-03 16:03

Yu Zhiquan

18-Dec-03 16:03

I run an application(e.g. Test.exe) in Debug or Release mode,
without manifest file, the applcation will good.

But, I met a GDI Objects leak issue,
when I run the application with
a manifest file(e.g. Test.exe.manifest).

the follow is in the manifest file:

<br />
     <dependency><br />
          <dependentAssembly><br />
               <assemblyIdentity<br />
                    type="win32"<br />
                    name="Microsoft.Windows.Common-Controls"<br />
                    version="6.0.0.0"<br />
                    publicKeyToken="6595b64144ccf1df"<br />
                    language="*"<br />
                    processorArchitecture="x86"<br />
               /><br />
          </dependentAssembly><br />
     </dependency><br />

The leak will not issue, when I remove the codes from the manifest file,
or, the leak will not issue , I rename the application name(e.g. ATest.exe).

The GDI Objects leak due to manifest file?
Who can give me some comments, thanks in advance!

Re: GDI Objects LEAK due to manifest file codes?

John R. Shaw18-Dec-03 16:50

dumbest question in the world

18-Dec-03 16:50

I have no idea, but the manifest from Microsoft doe not cause this problem:

<assembly xmlns="urn:schemas-microsoft-com:asm.v1" manifestversion="1.0">
<assemblyidentity
="" version="1.0.0.0" processorarchitecture="X86" name="CompanyName.ProductName.YourApp" type="win32">
<description>Your application description here.
<dependency>
<dependentassembly>
<assemblyidentity
="" type="win32" name="Microsoft.Windows.Common-Controls" version="6.0.0.0" processorarchitecture="X86" publickeytoken="6595b64144ccf1df" language="*">

INTP

rajdawg18-Dec-03 15:03

rajdawg

18-Dec-03 15:03

Why doesn't VC++ recognize:
#include <string>

I see this in all the books that I am reading, but when I use it...I get:
'string': undeclared identifier
on the line where I try to declare my strings.
Incidentally, Bloodshed's C++ compiler (Dev-C++ 4.9.8.0) doesn't recognize it either.

I also tried
#include <string.h>
#include<string>
#include<string.h>
and every other configuration I could think of, include using "" around 'string'
I now this is incredibally dumb...but there it is.
thanks! $Shucks | :-\$
roger

oops....lets try this with code tags

rajdawg18-Dec-03 15:05

rajdawg

18-Dec-03 15:05

<br />
Why doesn't VC++ recognize:<br />
#include <string><br />
<br />
I see this in all the books that I am reading, but when I use it...I get:<br />
'string': undeclared identifier<br />
on the line where I try to declare my strings.<br />
Incidentally, Bloodshed's C++ compiler (Dev-C++ 4.9.8.0) doesn't recognize it either.<br />
<br />
I also tried <br />
#include <string.h><br />
#include<string><br />
#include<string.h><br />

Re: oops....lets try this with code tags

rajdawg18-Dec-03 15:06

rajdawg

18-Dec-03 15:06

nevermind

Re: oops....lets try this with code tags

Christian Graus18-Dec-03 15:28

Christian Graus

18-Dec-03 15:28

You figured out how to use namespace std then ?

For future reference, you need to check 'Do not treat <'s as HTML tags' when posting stuff that uses < >.

Christian

I have drunk the cool-aid and found it wan and bitter. - Chris Maunder

Re: dumbest question in the world

Michael Dunn18-Dec-03 18:06

Michael Dunn

18-Dec-03 18:06

See the FAQ
2.1 I'm trying to use a standard C++ library class (like cout, cin, or string) but the compiler gives an undeclared identifier error (C2065) on those names. Why?[^]

--Mike--
Ericahist | CP SearchBar v2.0.2 | Homepage | RightClick-Encrypt | 1ClickPicGrabber
Pinky, are you pondering what I'm pondering?
I think so Brain, but if we shaved our heads, we'd look like weasels!

Application Error

help_me_pleeze18-Dec-03 15:01

help_me_pleeze

18-Dec-03 15:01

My application compiles and builds properly. there is no error.
But when i get the application interface and click on the menu
to open a dialog boxes which has the controls in it, rather than opening
i get this message.

-------------------------------------------------------

Blah Blah .....

Please tell microsoft about this problem.

Blah Blah .....

We have created an error report that you can sedn to us.
We will treat this report as confidential and anonymous.

To see what this error report contains, CLICK HERE.

When i click on "CLICK HERE" i get this message

--------------------------------------------------------------

I get this error first:
--------------------------------------------------
Error Signature------------------------------------
AppName: Application App Ver 2.0.0.2 ModName ntdll.dll
ModVer: 5.1.2600.1217 Offset: 000096b5

Reporting details------------------------------

Blah Blah

To view technical information about the error report, click here
---------------------------------------------------------------
Then on seeing the error and exiting i get this messagebox at the end.

|--------------------------------------------------|
|Application: Application.exe Application Error |
|------------------------------------------------- |
|The instruction at "0x77f58dc5"referenced memory |
|at "0xfffffff8". The memory could not be read. |
|Click OK to terminate the program. |
| |
| ----- |
| |OK | |
| ----- |
|--------------------------------------------------|

What is wrong wi th my application ?
pleae help me out i have been battling this for the last one week
with no progress.
Thanks.

.iCe. Confused | :confused:

Joe Woodbury18-Dec-03 15:22

Joe Woodbury

18-Dec-03 15:22

Odds are it means you passed an invalid pointer to something.

My immediate guess is that you have are using a dialog ID for which there is no control.

Anyone who thinks he has a better idea of what's good for people than people do is a swine.
- P.J. O'Rourke

secret_agent19-Dec-03 6:59

secret_agent

19-Dec-03 6:59

Ok here is a little history how i started getting errors.
I made some child dialogs boxes . After that attached them to the Parent Dialog Box. Everything is working fine. Then i added the control variables on the dialog boxes which are Check Boxes and Edit boxes only.
Now i am intialising them in the

DoDataExchange(CDataExchange* pDX)

Constructor.
e.g. setting control m_variable1 = 0.5
and so on.

After doing so it gives me error.

Any more suggestions.

Thankks for answering though.

============================================================================
A winner is not one who never fails, but one who NEVER QUITS!
============================================================================
7 days and i havent quit yet !

Joe Woodbury19-Dec-03 7:28

Joe Woodbury

19-Dec-03 7:28

Just check the variables and step through the code. Make sure everything is being initialized properly and in the right order. Make sure you aren't using any of the controls before they are created.

Good luck.

Anyone who thinks he has a better idea of what's good for people than people do is a swine.
- P.J. O'Rourke

John R. Shaw18-Dec-03 16:12

18-Dec-03 16:12

You may have passed an invalid pointer or you may be over writting memory in an earlier function call. The last time I saw this error I had to back track until I found a miscalulated value that was causing an over write. This error normaly appears after the cause of the error (some times long afer).

INTP