Refactor[BMQT]: make Subscription rule-of-three complaint #370
Conversation
| #if defined(BSLS_COMPILERFEATURES_SUPPORT_RVALUE_REFERENCES) && \ | ||
| defined(BSLS_COMPILERFEATURES_SUPPORT_NOEXCEPT) | ||
| /// Assign to this object the value of the specified `rhs` object. | ||
| /// After performing this action, the `rhs` object will be left in a | ||
| /// valid, but unspecified state. | ||
| Subscription& operator=(Subscription&& rhs); | ||
| #endif |
There was a problem hiding this comment.
We should probably use bslmf::MovableRef instead of preprocessor guards here
There was a problem hiding this comment.
On the other hand, what do you think about removing explicit copy constructor and assignment at all? It should deal with the warnings, and also we re-implemented the default one anyway. We don't use allocators for these classes, so we don't need to implement allocator-aware assignment.
There was a problem hiding this comment.
The proposed change is in the latest commit
There was a problem hiding this comment.
This would also work, for some reason I thought there was an explicit destructor here that was forcing us into a rule-of-3/rule-of-5 situation
Signed-off-by: Evgeny Malygin <emalygin@bloomberg.net>
678098
force-pushed
the
240723_cleanup_implicit_bmqt_Subscription
branch
from
ca40d2d to
e309272
Compare
There was a problem hiding this comment.
See https://bloomberg.github.io/bde-resources/doxygen/bde_api_prod/classbslmf_1_1MovableRef.html for how we can rewrite this to have rvalue references work on platforms with >=C++11 while also having fake moves on C++03
Signed-off-by: Evgeny Malygin <emalygin@bloomberg.net>
678098
changed the title
Removes ~400 lines from the build log
Compliant to https://en.cppreference.com/w/cpp/language/rule_of_three