|
Boost Users : |
From: Robert Ramey (ramey_at_[hidden])
Date: 2007-02-08 21:52:50
The documentation of 1.33 does describe issues related to header ordering.
Header order restrictions have been eliminated in 1.34.
Robert Ramey
Jeffrey Brent McBeth wrote:
On Thu, Feb 08, 2007 at 08:11:57AM -0800, Robert Ramey wrote:
> I see nothing wrong with your example.
>
> I compiled your code with no problems or messages of any kind on the
> following compilers:
> VC 7.1, Borland 5.51 Borland 5.64, Commeau, gcc 3.3
>
> I am using the Boost Head branch. But that shouldn't make a difference.
Wow, okay. I'm doing something seriously wrong then. I tried the same code
on my VC7.1 today and got the same error about incomplete types as on my Mac
and Linux boxen. I have one more Linux box with a compiler I can try it on
(it is a different distribution)
demo.cpp compiles and appears to work fine.
And, in fact if I add a std::string to it, it continues to work fine.
Hmmm... Ahhh, I haven't seen that in a while.
If I switch the archive includes with the serialization includes, it
compiles.
So,
#include <boost/archive/text_oarchive.hpp>
#include <boost/archive/text_iarchive.hpp>
#include <boost/serialization/string.hpp>
Rather than
#include <boost/serialization/string.hpp>
#include <boost/archive/text_oarchive.hpp>
#include <boost/archive/text_iarchive.hpp>
which does not compile.
That is going to be a pain. Thanks all for your patience. Is this header
ordering documented, did I hit a weird compiler bug on my systems, or am I
smoking rope?
Jeff.cgi/boost-users
Boost-users list run by williamkempf at hotmail.com, kalb at libertysoft.com, bjorn.karlsson at readsoft.com, gregod at cs.rpi.edu, wekempf at cox.net