|
Hi everybody,
I have a console application, that create a dialog box to show a progress bar. This has worked fine for years (it started as a Visual Studio 6 project, then I migrated it to Studio 2005) Now, I have migrated it to Studio 2010. It still works fine under Windows 7, but under XP, the dialog box returns -1, with a GetLastError() of 0. When I look at the messages sent to the dialog box, then I see WM_SETFONT, followed by WM_DESTROY and WM_NCDESTROY. I call it as
DialogBox (hInstance, MAKEINTRESOURCE (IDD_MAPPER), HWND_DESKTOP, DialogProc), where DialogProc is defined as
INT_PTR CALLBACK DialogProc(HWND, UINT, WPARAM, LPARAM);
Like I said: it all worked fine and still does in Windows 7, but XP does not seem to co-operate anymore.
Anybody any ideas? Would be grately appreciated!
Thanks in advance,
William
|
|
|
|
|
Is you Windows XP updated to (at least) Service Pack 2 ?
(see here here[^])
Veni, vidi, vici.
|
|
|
|
|
It is most likely the manifest. When I converted my projects to VS2010, it changed all my manifests to use urn:schemas-microsoft-com:asm.v2 in this line:
<assembly xmlns="urn:schemas-microsoft-com:asm.v1" manifestVersion="1.0>
You need to change it back to v1.
Basically, manifest should look similar to:
="1.0"="UTF-8"="yes"
<assembly xmlns="urn:schemas-microsoft-com:asm.v1" manifestVersion="1.0">
<assemblyIdentity
version="2.0.1.1"
processorArchitecture="X86"
name="myAwesomeApp"
type="win32"
/>
<description>Awesome</description>
<dependency>
<dependentAssembly>
<assemblyIdentity
type="win32"
name="Microsoft.Windows.Common-Controls"
version="6.0.0.0"
processorArchitecture="X86"
publicKeyToken="6595b64144ccf1df"
language="*"
/>
</dependentAssembly>
</dependency>
</assembly>
Also, please check if you still using the manifest loaded via RT_MANIFEST resource. You need to delete this from resource and use the project settings. VS2010 has a habit of assigning RT_MANIFEST to 2, making the program unusable under XP.
And of cause you need to check this linker option: MinimumRequiredVersion. It should be 5.00.
|
|
|
|
|
Thanks for your answer, but I cannot find much out of the ordinary. I am not very well familiar with the manifest, but it looks to be generated by Studio2010? Anyway, it reads:
<?xml version='1.0' encoding='UTF-8' standalone='yes'?>
<assembly xmlns='urn:schemas-microsoft-com:asm.v1' manifestVersion='1.0'>
<trustInfo xmlns="urn:schemas-microsoft-com:asm.v3">
<security>
<requestedPrivileges>
<requestedExecutionLevel level='asInvoker' uiAccess='false' />
</requestedPrivileges>
</security>
</trustInfo>
</assembly>
There is no RT_MANIFEST mentioned in the resource
Also, the MinimumRequiredVersion is left blank, so I don't expect any difficulties there!
In any case: the program itself runs and it does everything up untill the point that I start the dialog box. (and everything after that dialogbox crashes) There are a number of log statements before and after and I see everyone of them. Only everything that occurs from within the dialogbox doe not appear in XP.
|
|
|
|
|
Try to set MinimumRequiredVersion to 5.00
|
|
|
|
|
Just tried that: No change 
Thanks anyway
|
|
|
|
|
Is InitCommonControlsEx() of any help?
"One man's wage rise is another man's price increase." - Harold Wilson
"Fireproof doesn't mean the fire will never come. It means when the fire comes that you will be able to withstand it." - Michael Simmons
"Show me a community that obeys the Ten Commandments and I'll show you a less crowded prison system." - Anonymous
|
|
|
|
|
YES! Thank you very much! I had commented that one out, since Studio 2005 at the time said it was no longer necessary! Including it again did the trick!
Thank you very much!
|
|
|
|
|
hi,
I study ado program ,when DB restart, normally use _ConnectionPtr's Open to reConnect firstly; but if I don't do it, _RecordsetPtr also can work...why?
|
|
|
|
|
Hi
I did code to draw curves using PolyBezier function. now I want to allow user to draw it using the mouse but I do not know how I can do it how I specify the control points for the curve while user draw the curves using mouse??
can you help me in doing this did lots of search without any result
thanks
|
|
|
|
|
It depends on what you want. You could, for instance, let the user draw the curve, extrapolate control points from the drawn curve (aka x,y positions that the mouse went through), and then use them to calculate a smooth curve. Or you could have the user draw a line, then click a point on that line to create a control point and allow them to move that point. I've seen both.
|
|
|
|
|
Hi,
I have created a COM dll . I'm calling another dll(C#.net dll\assembly)by adding refrence of this .net dll(from project properties) .And i able to access properly.
Now i want to make this .net dll usage should be independent of path. I mean -
I'll not add any reference or give path name in my COM dll Hence my COM dll don't know where this .net dll is located in the system. While COM Dll try to call this .net dll it'll search in whole system & load it.
Kindly please advice the C++ code for this
OR LoadLibrary() is only the method through that only we can load library dynamically(independent of path)
Please advice . Thanks in advance 
modified 6-Nov-12 8:05am.
|
|
|
|
|
litu kumar wrote: it'll search in whole system
That is unlikely to be a good idea.
Have you tried to write code that searches for one dll on the system?
What if there are two different versions?
What if there are mapped network drives?
|
|
|
|
|
Hi jschell,
You are right - its not a good idea.
We don't want to give specific location of dependent Dll in COM code . Because this dependent DLL is not developed by us. This dll usually obtained after installation of a software.
we legally tied-off with this software company & code cann't be provided them(as usual company policy)
Hence if it is possible then we want to load dynamically.
I'm guessing if we do search the dependent dll in whole system while invoking ,there may be the performance issue may happen. But still the requirement is like this.
Hope you understood our requirement.
Hence if any best ways kindly please suggest.
modified 7-Nov-12 0:10am.
|
|
|
|
|
So you want to detect if it is installed.
Normal installations will create a registry entry which you can look for. You find the entry and grab the install directory. Then...
During your install.
1. Present a dialog to the user with the default product path (or the found one above.)
2. The user can modify the path.
3. When the user says ok you then verify that the correct dll is at the path specified.
4. If the dll (and correct version) is not found then put out an error message and go back to step 1.
|
|
|
|
|
I've been using the auto keyword, lambdas and move constructors so far and all seems very cool. If the only thing they added was the auto keyword it would still be worth upgrading. This is just awesome...
auto it = someCollection.begin();
...but of course the downside is we now need cbegin() as well
|
|
|
|
|
This might help in lot of cases, but I really wonder if C++ 11 can reduce the pain of memory mangement. In my experience, auto variables are hardly I needed for my development. but really got tired of several gotchas and memory corruption and management with C++
-Sarath.
Rate the answers and close your posts if it's answered
|
|
|
|
|
_Josh_ wrote: auto it = someCollection.begin();
[puts on curmudgeonly old git voice] A total abomination which will lead to lazy programmers (as if there weren't enough already), and unreadable code.
One of these days I'm going to think of a really clever signature.
|
|
|
|
|
Don't get me started. Today's object spaghetti makes the old C spaghetti pale in comparison.
C++11 advocate: Look, I made the code more awesome.
Me: It's slower, more complicated than it needs to be and much harder to find bugs. Why not just use C#?
C++11 advocate: But, it's more awesome!
(I actually heard someone today advocate that all pointers must be wrapped in shared_ptr, no matter what. I'm even working on some code that did that for absolutely no apparent reason--the delete happens within three lines and there is a catch around the call to which the pointer is passed AND the pointer is not actually shared. Yet, they not only used a shared_ptr, but spent time creating a factory to instantiate the only instantiation ever of the object to which the pointer is passed.)
|
|
|
|
|
But the variable is still typed. There is not much in c++ that can't be abused. I think used well auto will make life better.
|
|
|
|
|
_Josh_ wrote: I think used well auto will make life better.
I don't see how writing auto rather than a type will make life better.
One of these days I'm going to think of a really clever signature.
|
|
|
|
|
Just easier. At first I was quite sceptical, anything that looks like it might break c++'s type safety is bound to raise suspicion but after reading and playing with it a bit it's actually harder to shoot yourself in the foot than I'd first thought.
The rvalue references are a much bigger change but we've seen some good performance benifits from that.
|
|
|
|
|
you've come a long way, "void *"
|
|
|
|
|
Indeed, a var created with auto is typed. void * is still awesome occasionally.
|
|
|
|
|
Hasn't auto always been a part of the C language? Maybe it has a new meaning these days?
"One man's wage rise is another man's price increase." - Harold Wilson
"Fireproof doesn't mean the fire will never come. It means when the fire comes that you will be able to withstand it." - Michael Simmons
"Show me a community that obeys the Ten Commandments and I'll show you a less crowded prison system." - Anonymous
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
