Re: [Libreoffice-commits] core.git: Branch 'libreoffice-5-0' - download.lst external/libetonyek

Hi,

On Sat, Jun 27, 2015 at 12:27:59AM +0200, Thorsten Behrens wrote:
> David Tardon wrote:
> > commit 60e38140db23ce714b172f8e60d71da4e4c495e5
> > Author: David Tardon <dtardon@redhat.com>
> > Date:   Tue Jun 2 17:49:48 2015 +0200
> >
> >     upload libetonyek 0.1.3
> >     
> >     (cherry picked from commit ee3d40b30816a8fc6d4e8f984659c8dfac19ec3b)
> >     
> >     Change-Id: I451f02ee93a798c9c1712082d59a7ff96b0d4a4e
> >     Reviewed-on: https://gerrit.libreoffice.org/16508
> >     Reviewed-by: Michael Stahl <mstahl@redhat.com>
> >     Tested-by: Michael Stahl <mstahl@redhat.com>
>
> Hi David,
>
> sadly that breaks on gcc4.6:
>
> configure:17296: result: yes
> configure:17318: checking for MDDS
> configure:17403: result: yes
> configure:17407: checking checking if mdds::flat_segment_tree can store values of any type
> configure:17419: /usr/bin/ccache /usr/bin/g++-4.6 -c -g -O2  
> -I/src/libo_tb-5-0/workdir/UnpackedTarball/mdds/include  
> -I/src/libo_tb-5-0/workdir/UnpackedTarball/boost conftest.cpp >&5
> In file included from 
> /src/libo_tb-5-0/workdir/UnpackedTarball/mdds/include/mdds/flat_segment_tree.hpp:35:0,
>                  from conftest.cpp:45:
> /usr/include/c++/4.6/limits: In static member function 'static _Tp 
> std::numeric_limits<_Tp>::max() [with _Tp = Value]':
> /src/libo_tb-5-0/workdir/UnpackedTarball/mdds/include/mdds/flat_segment_tree_def.inl:49:5:   
> instantiated from 'mdds::flat_segment_tree<_Key, 
> _Value>::flat_segment_tree(mdds::flat_segment_tree<_Key, _Value>::key_type, 
> mdds::flat_segment_tree<_Key, _Value>::key_type, mdds::flat_segment_tree<_Key, 
> _Value>::value_type) [with _Key = int, _Value = Value, mdds::flat_segment_tree<_Key, 
> _Value>::key_type = int, mdds::flat_segment_tree<_Key, _Value>::value_type = Value]'
> conftest.cpp:47:55:   instantiated from here
> /usr/include/c++/4.6/limits:313:48: error: no matching function for call to 'Value::Value(int)'
> /usr/include/c++/4.6/limits:313:48: note: candidates are:
> conftest.cpp:46:8: note: Value::Value()
> conftest.cpp:46:8: note:   candidate expects 0 arguments, 1 provided
> conftest.cpp:46:8: note: Value::Value(const Value&)
> conftest.cpp:46:8: note:   no known conversion for argument 1 from 'int' to 'const Value&'
> configure:17419: $? = 1
> configure: failed program was:
> | /* confdefs.h */
>
> Any easy workaround I should try?

Yes, make mdds.clean . This is not a gcc problem--the test is for a
feature that was only added in mdds 0.12.1. All deps are there, so mdds
should have been rebuilt--I do not know why it weren't.

D.