BIND 10 #2356: configure fails in examples/ because it doesn't find the boost header

BIND 10 Development do-not-reply at isc.org
Fri Nov 2 18:20:14 UTC 2012


#2356: configure fails in examples/ because it doesn't find the boost header
-------------------------------------+-------------------------------------
                   Reporter:  jreed  |                 Owner:  jinmei
                       Type:         |                Status:  reviewing
  defect                             |             Milestone:
                   Priority:         |  Sprint-20121106
  medium                             |            Resolution:
                  Component:  build  |             Sensitive:  0
  system                             |           Sub-Project:  DNS
                   Keywords:         |  Estimated Difficulty:  3
            Defect Severity:  N/A    |           Total Hours:  0
Feature Depending on Ticket:         |
        Add Hours to Ticket:  0      |
                  Internal?:  0      |
-------------------------------------+-------------------------------------

Comment (by jinmei):

 Replying to [comment:11 jreed]:

 Thanks for the review.

 > > {{{
 > > g++  -g -O2 -L/home/vorner/testing/bind10/lib/
 -R/home/vorner/testing/bind10/lib/ -o b10-host host.o -lb10-dns++
 > > g++: error: unrecognized option ‘-R’
 > > }}}
 >
 > Should try -Wl,-R  I think.  That is tried first in the test.

 Initially I didn't understand why it failed, because the check is
 basically a verbatim copy from the original configure.ac.  I then
 found several issues:

 - "-R" was actually hardcoded in ax_isc_bind10.m4.  This is a clear
   bug and was fixed.
 - I'm not sure if that was your case, but I've noticed the position of
   LDFLAGS is different when used in AC_TRY_LINK and in the Makefile
   (generated by automake).  Some version of g++ only complains about
   the latter case.  So it's possible that AC_TRY_LINK thinks -R is
   usable but it causes build time failure.  So, to make it safer, I
   used CXXFLAGS instead of LDFLAGS in the m4 check (and also set
   CCFLAGS in case the main script uses C).  And, it may still not be a
   complete solution because yet other versions of g++ only complains
   about it but doesn't treat it as an error.  So, as a final defense
   I actually changed the order of the check: -Wl,-R first and then -R.
 - I've updated the result message when -R is chosen with a note about
   the possible issue - even though it's less likely people actually
   read it.

-- 
Ticket URL: <http://bind10.isc.org/ticket/2356#comment:12>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development


More information about the bind10-tickets mailing list