Internal Compiler Error using temporary object containing gcroot

[Vladimir Kouznetsov]

#using <mscorlib.dll >
#include <vcclr.h>
public __gc class A {};
struct B {
void f(A* e);
gcroot<A*> e;
};
int main() {
B().f(new A()); // note that B b; b.f(new A)); works fine
return 0;
}

[Vladimir Kouznetsov]

Oh, well.

Either everybody knows or nobody cares...
"Vladimir Kouznetsov" <vl************ *****@ngrain.co m> wrote in message
news:eG******** ******@TK2MSFTN GP09.phx.gbl...

#using <mscorlib.dll >
#include <vcclr.h>
public __gc class A {};
struct B {
void f(A* e);
gcroot<A*> e;
};
int main() {
B().f(new A()); // note that B b; b.f(new A)); works fine
return 0;
}

[David Lowndes]

>Oh, well.

Either everybody knows or nobody cares...

FWIW, the ICE happens with the alpha release Whidbey compiler too -
I'll try to report it to MS.

Dave
--
MVP VC++ FAQ: http://www.mvps.org/vcfaq

[David Lowndes]

>#using <mscorlib.dll >

#include <vcclr.h>
public __gc class A {};
struct B {
void f(A* e);
gcroot<A*> e;
};
int main() {
B().f(new A()); // note that B b; b.f(new A)); works fine
return 0;
}

While it's not correct that the compiler gives an ICE, is the code
example valid?

Dave
--
MVP VC++ FAQ: http://www.mvps.org/vcfaq

[Vladimir Kouznetsov]

Why would you think it's not?

thanks,
v

"David Lowndes" <da****@mvps.or g> wrote in message
news:vk******** *************** *********@4ax.c om...

#using <mscorlib.dll >
#include <vcclr.h>
public __gc class A {};
struct B {
void f(A* e);
gcroot<A*> e;
};
int main() {
B().f(new A()); // note that B b; b.f(new A)); works fine
return 0;
}

While it's not correct that the compiler gives an ICE, is the code
example valid?

Dave
--
MVP VC++ FAQ: http://www.mvps.org/vcfaq

[David Lowndes]

>Why would you think it's not?

Because I don't use managed C++ myself and I'm really not sure what's
right, wrong, or indifferent with it!

Dave
--
MVP VC++ FAQ: http://www.mvps.org/vcfaq

[Vladimir Kouznetsov]

Thank you for your feedback. Sadly I expected someone from MS reacted to
that. Next time I most probably won't report.
TTBOMK this code is legal managed C++.

thanks,
v

"David Lowndes" <da****@mvps.or g> wrote in message
news:5v******** *************** *********@4ax.c om...

Why would you think it's not?

Because I don't use managed C++ myself and I'm really not sure what's
right, wrong, or indifferent with it!

Dave
--
MVP VC++ FAQ: http://www.mvps.org/vcfaq

[David Lowndes]

>Thank you for your feedback. Sadly I expected someone from MS reacted to

that. Next time I most probably won't report.
TTBOMK this code is legal managed C++.

I'm sorry that you seem disappointed that I replied. I replied because
no-one else had, and I was only being honest in admitting my lack of
knowledge in managed C++ usage.

FWIW, I had submitted your code as a bug report to MS.

Dave
--
MVP VC++ FAQ: http://www.mvps.org/vcfaq

[Vladimir Kouznetsov]

On no David, I'm grateful that you answered. I was disappointed that I did
not receive any "official" reaction. Now, knowing that it was properly
reported, I'm completely satisfied.
Thanks again!

v

"David Lowndes" <da****@mvps.or g> wrote in message
news:ap******** *************** *********@4ax.c om...

Thank you for your feedback. Sadly I expected someone from MS reacted to
that. Next time I most probably won't report.
TTBOMK this code is legal managed C++.

I'm sorry that you seem disappointed that I replied. I replied because
no-one else had, and I was only being honest in admitting my lack of
knowledge in managed C++ usage.

FWIW, I had submitted your code as a bug report to MS.

Dave
--
MVP VC++ FAQ: http://www.mvps.org/vcfaq

[Ronald Laeremans [MSFT]]

In general we don't post in threads where an MVP already addressed the issue
and in this case probably no one noticed that David said he didn't know
whether the example was valid code.

Ronald Laeremans
Group Program Manager
Visual C++ team

"Vladimir Kouznetsov" <vl************ *****@ngrain.co m> wrote in message
news:eq******** *****@TK2MSFTNG P11.phx.gbl...

On no David, I'm grateful that you answered. I was disappointed that I did
not receive any "official" reaction. Now, knowing that it was properly
reported, I'm completely satisfied.
Thanks again!

v

"David Lowndes" <da****@mvps.or g> wrote in message
news:ap******** *************** *********@4ax.c om...

Thank you for your feedback. Sadly I expected someone from MS reacted tothat. Next time I most probably won't report.
TTBOMK this code is legal managed C++.

I'm sorry that you seem disappointed that I replied. I replied because
no-one else had, and I was only being honest in admitting my lack of
knowledge in managed C++ usage.

FWIW, I had submitted your code as a bug report to MS.

Dave
--
MVP VC++ FAQ: http://www.mvps.org/vcfaq