[Nfd-dev] ndn-cxx compiling problem

[Davide Pesavento]

Yes, 1.59.0 is the version affected by the boost bug I was talking
about. As I said, we already added a workaround to the code[1], but
for some reason it's not working as expected.

[1] https://github.com/named-data/ndn-cxx/blob/master/src/util/logger.hpp#L118-L123

On Wed, Oct 19, 2016 at 7:36 PM, Weiwei Liu <summerwing10 at gmail.com> wrote:
> My boost version is 1.59.0. Does the bug exist in this version and what is
> the workaround?
>
>
> Thanks,
>
> Weiwei
>
>
>
> 2016-10-19 3:48 GMT-07:00 Davide Pesavento <davide.pesavento at lip6.fr>:
>>
>> On Wed, Oct 19, 2016 at 8:50 AM, Weiwei Liu <summerwing10 at gmail.com>
>> wrote:
>> > Dear folks,
>> >
>> > I'm getting errors when compiling ndn-cxx with tests. Does anyone know
>> > what
>> > is causing this problem?
>> >
>> > Here is some detailed information:
>> >
>> > ndn-cxx 010f0868cd204f75f661acc4320803d783786213
>> > OSX 10.10.5 LLVM version 7.0.2 (clang-700.1.81)
>> >
>> > [170/304] Compiling tests/unit-tests/util/io.t.cpp
>> >
>> > [171/304] Compiling tests/unit-tests/util/log-module1.cpp
>> >
>> > [172/304] Compiling tests/unit-tests/util/notification-stream.t.cpp
>> >
>> > ../tests/unit-tests/util/log-module1.cpp:33:3: error: invalid operands
>> > to
>> > binary expression ('basic_record_ostream<char_type>' and
>> > '::ndn::util::LoggerTimestamp')
>> >
>> >   NDN_LOG_TRACE("trace" << 1);
>> >
>> >   ^~~~~~~~~~~~~~~~~~~~~~~~~~~
>> >
>> > /Users/weiweiliu/development/ndn-cxx/src/util/logger.hpp:137:35: note:
>> > expanded from macro 'NDN_LOG_TRACE'
>> >
>> > #define NDN_LOG_TRACE(expression) NDN_LOG(TRACE, TRACE, expression)
>> >
>> >                                   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> >
>> > /Users/weiweiliu/development/ndn-cxx/src/util/logger.hpp:128:40: note:
>> > expanded from macro 'NDN_LOG'
>> >
>> >       NDN_BOOST_LOG(getNdnCxxLogger()) << ::ndn::util::LoggerTimestamp{}
>> > \
>> >
>> >       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ^  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> >
>> > ... ...
>> >
>>
>> We should already have a workaround for this in place (it's a boost
>> bug), or maybe it's a different issue.
>> What's your boost version?
>>
>> Thanks,
>> Davide
>
>