Boost Jamfile.v2 fixes

I've just committed the attached, to

1. Don't build impossible variants with "bjam --v2
install" on Linux.
2. Fix header install path, as reported by Roland Schwarz.

- Volodya


_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost-build

On Saturday 21 October 2006 14:32, Vladimir Prus wrote:
> I've just committed the attached, to
>
> 1. Don't build impossible variants with "bjam --v2
install" on Linux.
> 2. Fix header install path, as reported by Roland
Schwarz.

I've also committed this. The set of variants built on Linux
with V2 now is 
the same as with V1. The 'unversioned' libraries and
symlinks are not exactly 
right, though.

-  Volodya
_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost-build

Vladimir Prus wrote:
> I've also committed this. The set of variants built on
Linux with V2 now is 
> the same as with V1.

This is the problem I keep bringing up... It works for Linux
but doesn't 
for anything else. The set of libraries on Windows will now
be different 
from what BBv1 generates.



-- 
-- Grafik - Don't Assume Anything
-- Redshift Software, Inc. - http://redshift-software
.com
-- rrivera/acm.org - grafik/redshift-software.com
-- 102708583/icq - grafikrobot/aim - grafikrobot/yahoo
_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost-build

On Saturday 21 October 2006 20:11, Rene Rivera wrote:
> Vladimir Prus wrote:
> > I've also committed this. The set of variants
built on Linux with V2 now
> > is the same as with V1.
>
> This is the problem I keep bringing up... It works for
Linux but doesn't
> for anything else. The set of libraries on Windows will
now be different
> from what BBv1 generates.

Am I right the V1 produces different set of libraries on
Windows and Linux 
too?

If so, can we use "if" on OS name in V2?

- Volodya

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost-build

Vladimir Prus wrote:
> On Saturday 21 October 2006 20:11, Rene Rivera wrote:
>> Vladimir Prus wrote:
>>> I've also committed this. The set of variants
built on Linux with V2 now
>>> is the same as with V1.
>> This is the problem I keep bringing up... It works
for Linux but doesn't
>> for anything else. The set of libraries on Windows
will now be different
>> from what BBv1 generates.
> 
> Am I right the V1 produces different set of libraries
on Windows and Linux 
> too?

Yes, that's correct. More accurately different toolsets
produce only 
what they can 

> If so, can we use "if" on OS name in V2?

Well we could. But that may not account for differences in
toolsets 
within one OS. After all there are many toolsets (and C++
runtimes) on 
both Linux and Windows individually.


-- 
-- Grafik - Don't Assume Anything
-- Redshift Software, Inc. - http://redshift-software
.com
-- rrivera/acm.org - grafik/redshift-software.com
-- 102708583/icq - grafikrobot/aim - grafikrobot/yahoo

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost-build

On Saturday 21 October 2006 20:43, Rene Rivera wrote:
> Vladimir Prus wrote:
> > On Saturday 21 October 2006 20:11, Rene Rivera
wrote:
> >> Vladimir Prus wrote:
> >>> I've also committed this. The set of
variants built on Linux with V2
> >>> now is the same as with V1.
> >>
> >> This is the problem I keep bringing up... It
works for Linux but doesn't
> >> for anything else. The set of libraries on
Windows will now be different
> >> from what BBv1 generates.
> >
> > Am I right the V1 produces different set of
libraries on Windows and
> > Linux too?
>
> Yes, that's correct. More accurately different toolsets
produce only
> what they can 
>
> > If so, can we use "if" on OS name in V2?
>
> Well we could. But that may not account for differences
in toolsets
> within one OS. After all there are many toolsets (and
C++ runtimes) on
> both Linux and Windows individually.

Well, let's look at my changes.

1. Building both <link>static and <link>shared.
I think this is just right.
2. Building without <runtime-debugging>on. It looks
like V1 does not request 
this variant, right?
3. I've also removed the <runtime-link>static /
<runtime-link>shared. In 
retrospect, I'm not sure why it caused problem -- the static
runtime should 
have being just ignored. I'll look again.

- Volodya






_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost-build

> I've just committed the attached, to
>
> 1. Don't build impossible variants with "bjam --v2
install" on Linux.

I just compiled CVS HEAD (gcc on linux) with v2 and noticed
that no static
libraries are built.  Is this expected?  Up to now I've
built boost with
v1 and have used the static version of the
unit_test_framework...

Chris
_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost-build