Boost logo

Boost-Build :

From: Rene Rivera (grafik.list_at_[hidden])
Date: 2005-02-11 17:53:29

Robert Ramey wrote:
> I'm confused about the role of <include> and <sysinclude> in a V1 Jamfile.
> I don't see it explained anywhere.

AFAIK it's only explained in features.jam where they are declared..

# Add a user include path (#include "x.h").
path-feature "include" ; # <include>../foo/bar/mumble

# Add a system include path (#include <x.h>).
path-feature sysinclude ; # <sysinclude>/fu/man/chu

Whether the toolset can honor that completely is another story :-)

> Some build Jamfile include under requiremens
> <include>$(BOOST_ROOT)
> others include
> <sysinclude>$(BOOST_ROOT)
> at least one include both.

Ideally everyone should be using <sysinclude>$(BOOST_ROOT)

> The impression I had was that this is automatically included from
> boost-base.jam or ?

Do you mean that a <sysinclude>$(BOOST_ROOT) gets done automatically for
all? No. Although there are some toolsets which add the path directly
for some unknown reason. And a <sysinclude>$(BOOST_ROOT) gets added to
all test targets automatically.

> Note the three C:\BoostMainCVS in a row.

That should not be a problem, it is wasteful :-)

>This suggests to me that maybe I
> haven't specified include paths properly.

>Also the error message suggests
> that I'm missing some path or environmental variable or who knows what. I
> don't see any includes for como/libcomo or as I expect.

That usually only happens when something like COMO_PATH, or a related
var, is not configured.

> Note that this worked before, I updated the my local CVS tree and I started
> to get this. I also notice recent updates in tools/build/v1
> (allyourbase.jam - 3 weeks ago and boost-base.jam 8 days ago) whose log
> entries suggest that perhaps they might be related to the problems I'm
> experiencing.

Those would not explain the absence of the como include and lib paths.

> I also notice that CW and gcc-2.95-stlport builds in the test matrix are
> generating what seem to me similar errors - though with the serialization
> library build rather than the test library build.
> I'm don't know how to address this. Could someone check this for me?

I'll see if I can replicate it with cw-8_3 as that's what I have.

-- Grafik - Don't Assume Anything
-- Redshift Software, Inc. -
-- rrivera/ - grafik/ - 102708583/icq

Boost-Build list run by bdawes at, david.abrahams at, gregod at, cpdaniel at, john at