From: Beman Dawes (bdawes_at_[hidden])
Date: 2002-07-23 08:55:23
At 08:30 AM 7/23/2002, David Abrahams wrote:
>----- Original Message -----
>From: "Beman Dawes" <bdawes_at_[hidden]>
>> Why is the boost root ever specified as what GCC calls a "non-system"
>> directory? Since the standard specifies that a "..." include which
>> is then treated as a <...> include, it wouldn't seem necessary to
>> the boost root as a "non-system" directory.
>The culprit is this line (allyourbase.jam:1028):
> HDRS on $(<) = [ unique $(LOCATE_SOURCE) $(SEARCH_SOURCE) $(HDRS) ] ;
>which adds SEARCH_SOURCE to HDRS. You might try
> HDRS on $(<) = [ unique $(LOCATE_SOURCE) $(HDRS) ] ;
>instead. I can't promise this won't destabilize something. If there's a
>problem for the Python library, I can try to work around it.
That seems to fix the problem, but I only did some light testing with a few
compilers. I didn't try -a to force an entire retest of everything,
although I did -a on any_test for all the compilers I have.
Should I commit the change to CVS?
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