[Debian-med-packaging] Bug#914049: Fix

[Giovanni Mascellani]

Hi,

this is fortunately relatively easy: the push_back is fine, the problem
is the argument. Basically the argument passed to the push_back is a
pointer, but the vector to which we want to push is a vector of tuples,
whose first entry is actually an argument. That is fine, because a Boost
tuple can be constructed from its first entry (the other entries will
simply be default-constructed).

The point is that up to Boost 1.62 the tuple constructor constructing
from the first entry was not marked as "explicit", so the C++ compiler
could just invoke it implicitly. Now the constructor has become
"explicit", so it must be called explicitly. See the attached patch
("NeighborhoodGroup" is an alias for the tuple type).

Giovanni.
-- 
Giovanni Mascellani <g.mascellani at gmail.com>
Postdoc researcher - Université Libre de Bruxelles
-------------- next part --------------
Author: Giovanni Mascellani <gio at debian.org>
Bug-Debian: https://bugs.debian.org/914049
Subject: Fix FTBFS with boost1.67

Make an explicit call to the constructor for NeighborhoodGroup (which is
an alias for a specialization of boost::tuple), which is not marked
as "explicit".

--- progressivemauve-1.2.0+4713+dfsg.orig/src/repeatoire.cpp
+++ progressivemauve-1.2.0+4713+dfsg/src/repeatoire.cpp
@@ -1591,7 +1591,7 @@ void processNovelSubsetMatches( GappedMa
 		getSubsets(M_j,-direction*ni_parity*nj_parity).push_back(nj_link);
         //getExtraSubsets(M_j,-direction*ni_parity*nj_parity).push_back(nj_link);
 		// push M_n onto the heap
-		novel_subset_list.push_back(M_n);
+		novel_subset_list.push_back(NeighborhoodGroup{M_n});
 		//procrastination_queue.push(M_n);
 		novel_subset_count++;
 	}
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 228 bytes
Desc: OpenPGP digital signature
URL: <http://alioth-lists.debian.net/pipermail/debian-med-packaging/attachments/20181201/abb27376/attachment-0001.sig>