Re: [boost] [context] [windows] Intel C++ and GCC Support?

Subject: Re: [boost] [context] [windows] Intel C++ and GCC Support?
From: Joshua Boyce (raptorfactor_at_[hidden])
Date: 2012-04-22 04:16:49

On Sun, Apr 22, 2012 at 1:45 AM, Lars Viklund <zao_at_[hidden]> wrote:

> On Sat, Apr 21, 2012 at 05:00:23PM +0200, Oliver Kowalke wrote:
> > Am 21.04.2012 16:30, schrieb Joshua Boyce:
> >> Unfortunately I can't use the MSVC command line shell with Intel, as
> >> Boost.Build won't work with Intel unless it's run under the Intel
> >> shell (and even then it's kinda flaky)...
> >
> > strange - with mingw and MSVC command shell it works
>
> The intel toolchain has some strange file association madness going on,
> where it's unusable outside of the environment that the command prompt
> produces.
>
> It might be worthwhile investigating what the command prompt does and
> reproduce it in Boost.Build, so that you don't have the strange
> environmental dependency that BB tends to want to avoid.
>
> The typical fault you get is a Windows message box popping up asking
> what program to open a 'rsp' file with, or something... it's been a
> couple of years since I tried.
>
> --
> Lars Viklund | zao_at_[hidden]
>
> _______________________________________________
> Unsubscribe & other changes:
> http://lists.boost.org/mailman/listinfo.cgi/boost
>

Yeah, this is exactly what's going on. Boost.Build is also passing at least
one non-existent command line argument (/favor:blend) and it is trying to
run the iclvars.bat script (or w/e it's called) in the incorrect location
(this is the root of the problem I believe, but I haven't gotten around to
fixing it since I started porting to intel a few days ago as it's only a
2nd tier compiler for me).