BOOST_LEAF_CHECK and extra semicolons

[Zachary Michaels]

BOOST_LEAF_ASSIGN has no enclosing block, but BOOST_LEAF_CHECK does. This seems inconsistent and means that BOOST_LEAF_CHECK(mycoolthing()); results in an extra semicolon (and potentially a warning).

[Emil Dotchevski]

Are you thinking

#define BOOST_LEAF_CHECK(r)\
    static_assert(::boost::leaf::is_result_type<typename std::decay<decltype(r)>::type>::value, "BOOST_LEAF_CHECK requires a result type");\
    auto && BOOST_LEAF_TMP = r;\
    if( BOOST_LEAF_TMP )\
        ;\
    else\
        return BOOST_LEAF_TMP.error()

?

[Zachary Michaels]

Yes, something along those lines, I think.

I had

#define LEAF_CHECK(r)\
  static_assert(::boost::leaf::is_result_type<typename std::decay<decltype(r)>::type>::value, "BOOST_LEAF_CHECK requires a result type");\
  auto &&BOOST_LEAF_TMP = r;\
  if (!BOOST_LEAF_TMP) return BOOST_LEAF_TMP.error()

[Emil Dotchevski]

Thats not safe in a macro, the if is lacking an else.

[Zachary Michaels]

Makes sense!

[Zachary Michaels]

Actually, I spoke too soon. My macro-fu is weak. Would you just be worried about someone abusing it like LEAF_CHECK(...); else ...?

The one you proposed would be fine for me, if you think that one's acceptable.

[Emil Dotchevski]

Yes, exactly -- it is a very bad idea to leave an incomplete if statement in a macro. This was the main motivation for adding the scope actually, but it's true that it better without the scope, it would allow you to look at the temp variable during debugging.

[Zachary Michaels]

I see. Interesting to know the history, and thanks for the fix!