From: Reece Dunn (msclrhd_at_[hidden])
Date: 2005-12-05 11:15:11
Vladimir Prus wrote:
>On Monday 05 December 2005 12:33, Reece Dunn wrote:
> > Did you get the e-mail I sent you? I sent it a week or so ago, but
> > haven't received a reply.
>Yes, sorry, that dropped on the floor.
Oh well, never mind.
>Ok, speaking about dependent2, in client/Jamfile you have
> lib client : .. ../core ;
>if I replace this with
> lib client : ... ../core//core ;
>The error is gone.
>The bug there is that in:
> pch stdafx : [ cast _ pcheader : stdafx.h ] stdafx.cpp ;
>the hidded target created by "cast" is not marked as explicit, so when you
>refer to "../core", both "stdafx" and that hidden target get build. And so
>"client" target gets extra PCHHEADER target and generates another PCH file
>I'll fix this, but this is not problem with PCH implementation itself.
>Except -- why do you need "cast" in the first place?
I need it so that the PCH HPP file is not treated as a normal HPP file. The
issue is with generated headers (which is why the MC test is in there).
Otherwise, the build process gets confused with the:
MC ==> RC ==> HPP ==> PCH ==> ...
and you get an error. (I can't remember the details, see the PCH discussion
and my replies to Alexey for more information).
Thanks for the help :).
Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk