Thank you for the reply.
I believe my development environment is up to date.
I have a version of deque (v 2.33) which contains the fixes described at:
http://www.dinkumware.com/vc_fixes.html
The version of deque which generates the error contains the line (#258):
{_Construct(_F, _L, _Iter_cat(_F)); }
With the problem due to the call to
_Iter_cat(_F)
Specifically, _Iter_cat(_F) signature is: _Iter_cat(const_Ty *). We are providing a
value 1 for this parm when we call m_lengths(1,0).
I notice Crypto++ v5 initializes m_lengths with '(1,0U)' instead of (1,0).
Just to verify, could I trouble you to try compiling mqueue after changing line #9
from
m_lengths(1, 0U)
to
m_lengths(1, 0)?
If that is not the issue, then I must have another environment problem.
Thanks for your time.
*************
Perhaps you have STLport installed? If yes, you can work around it by
taking out the STLport dir when building the library. I never tried getting
it to work with STLport - maybe someone else has?
--
gl
----- Original Message -----
From: "Wei Dai" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Wednesday, November 06, 2002 1:14 AM
Subject: Re: Crypto++ 4.2 compile error - mqueue
> On Tue, Nov 05, 2002 at 07:02:06PM -0600, David Atkins wrote:
> > Apparently it does not like the initialization of m_lengths(1, 0) for
MessageQueue. Compilation is OK after removing the initialization.
>
> No, that initialization is necessary. I suspect something's wrong with
> your setup. Make sure you have service pack 5 installed. Also, try
> borrowing someone else's machine and compiling Crypto++ on it.
>
>
