Re: [lmi] building latest cvs with MSVC (PETE paths)

[Greg Chicares]

On 2008-11-17 17:53Z, Vaclav Slavik wrote:
> On Mon, 2008-11-17 at 02:35 +0000, Greg Chicares wrote:
>> Here are changes we could make, in increasing order of awfulness.
>> 
>> Not awful at all: change #include lines for 'et_vector.hpp'
>> in lmi files outside the PETE subdirectory. 
> 
> I like this the best -- it preserves PETE hierarchy and doesn't require
> PETE installation step (that's something I want to avoid if at all
> possible, the less unusual our Automake makefiles are, the better).
> 
> I may be missing something (another version of compiler, perhaps?), but
> the patch below works for me on both platforms. (I made sure to remove
> PETE from /opt/lmi/third_party. Also notice that the include directory
> is absolute, because $(src_dir) is /opt/lmi/src/lmi; that may be why it
> works.)

It works for me, too, with como and borland as well as gcc. And it's
a nice simplification. Thanks.

I used a q-char-sequence instead of an h-char-sequence [C++2003 2.8]
for including "et_vector.hpp", because elsewhere in lmi we have
  #include "header-in-lmi-hierarchy.hpp"
  #include </path/to/header-outside-lmi-hierarchy.h.or.hpp>
  #include <standard-header-with-no-extension>
, and
  #include "et_vector.hpp"
seemed more consistent with that existing practice. But let me know
whether this breaks anything for you.