First off, I have absolutely no experience with any form of version control software. However, if all you need is to make sure that a file in directory "A" is always identical to a file in directory "B" you can create an NTFS hardlink in "A" that points to the file in "B". For example, "\A\foolinked.h" would actually be the "\B\foo.h" file. Any use of "\A\foolinked.h" would always result in the indirect use of "\B\foo.h" with zero effort on your part (aside creating the initial link). Note that both files have to be on the same NTFS partition if I remember correctly. I believe there is a utility on one of the Windows CD's to do this (i.e. resource kit?). If not, I could send you some source to do this if you want to create your own command-line program. Hope this helps.

I asked her the following: Q: "How big are your mammary glands?" A: "Please, do not lower the tone of this conversation! Why don't you ask me some questions about our latest range of communication devices instead?" Q: "Do you like to swallow?" A: "To tell you the truth I have't really made up my mind yet." I'm laughing so hard I can barely breathe :laugh:

Lately, Tool's "Lateralis" CD. I wasn't too wild about it when I first bought it but now I can't stop listening to it. Another frequent play is Pink Floyd's "Wish You Were Here". Very calming and dreamy :cool:

I have just such a solution. Give me a week to clean up the code and make a test application showing how to use it.

My approach is to use bool everywhere remotely possible and to stay as far away from MFC as much as possible :). You never know when you might have to port your code or when MS might change their minds on BOOL. I just use conditional returns like this if a BOOL is required: if ( true == bState ) { return TRUE; } else { return FALSE; } when I have to return a BOOL in an overridden function. I try to never mix BOOL and bool. Even though they serve the same purpose, they are different types and should be treated as such.

For a CArray you have to dereference the internal data pointer to the C style array within the CArray. Any collection not using standard C style arrays for data storage, which I believe is everything but CArray derivatives, will not likely have its contents aligned in a contiguous portion of memory like a CArray will. So viewing such a container's contents would probably be very difficult if not impossible. However, viewing a CArray is easy. For example, if you are stepping through a class with a CArray derived data member named m_array you can see the contents of index 4 by entering this in the watch window: *m_array.m_pData[ 4 ] This assumes that that the array has a size of at least 5. Overstepping the allocated size of the array will probably not hurt in this situation but you will see garbage instead of what you expect to see. Just think of the watch window as a way to execute a limited line of code. Virtually any syntax you would use in your program should be valid in the watch window. However, don't forget about the scope of the data you want to watch. It's easy to assume that what you put in should work and then believe that the problem is not yours when it doesn't work as anticipated.

You said that you were hoping VCPP SP4 would fix the problem, did it?

I also encounter IDE pauses frequently but they are usually no longer than 2 or 3 seconds. You didn't say what kind of system you had, buy my system is a 800 MHz P3 it may be fast enough to make the pause less noticible. You are probably right about the delays being caused by intellisense. I thought about turning it off until it dawned on me how much I use it

Has anyone else noticed new problems with VC++ 6 after moving from NT4 to W2K? I got a new system with W2K and after installing VC++ 6 the IDE now has a few small but very annoying problems that I don't recall from the NT4 days. I haven't yet applied SP4 because of MS' terrible track record with service packs. I thought I'd wait a few months to see if there wer any problems. Anyway, the problems are: 1. Cut and paste of code doesn't always work. You can select the text then hit CTRL-C but the IDE WILL NOT paste it into the new location no matter how many times I try. Repeatedly re-selecting and re-copying doesn't do anything. However, if I change my copy selection just by one character it works just fine! 2. When trying to save a modified source file, which is part of a loaded project, the IDE will tell me that the file cannot be saved (I forget the exact error). But if I try again, almost immediately, the file saves just fine. It's like there are two threads working with the file but they don't have any cooperation. 3. The IDE will periodically crash when I hit F7 to rebuild. This was so rare under NT4 that I was really surprised when it began happening under W2K. 4. After doing a cut/copy or paste the IDE gets stuck in selection mode. That is, the last location of the carat becomes the start of a text selection and wherever you move the carat to becomes the end of the selection. No amount of ESC, CTRL-C or anything will get it out of this selection mode. I got really pissed and wiped my hands over the entire keyboard several times, like you would a piano keys, and the all went back to normal. That is now how I fix it. It sounds like I'm beating the shit out of my keyboard :)!!!

Like Nick, I've found that deleting *.ncb files seems to cure almost every problem I have with the IDE. I don't use any VCS system yet but I make zip backups of my source directories at least once a day. I've got the batch file that does the zipping setup to delete any .ncb files, among others, before it makes the archive. Be aware that the *.ncb file is locked while the project is loaded in the IDE. I seem to recall that the IDE occasionally fails to release the file lock when you close a project from the File menu. I now just exit the IDE to make sure that everything is closed properly.