Support nb::init(<lambda>) as syntactic sugar for custom constructors #885
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Unlike the corresponding pybind11 feature, the lambda must return by value, not by pointer. This simplifies the plumbing quite a bit and allows us to typically take advantage of guaranteed copy elision, producing basically the same code that we would have gotten from the previously-recommended manual
__init__
approach. Manual__init__
is still the most appropriate method if some action needs to be taken after calling the C++ constructor.This PR fully supports alias/trampoline classes, including the option to pass separate factory functions for constructing the bound type vs the trampoline type, except that I didn't bother providing an equivalent of pybind11's
init_alias
for forcing to always construct the trampoline type (since we don't have that for C++-provided constructors either). If this PR is judged too complex, the easiest way to simplify it would be by dropping support for trampoline classes. I'm a bit reluctant to do that though, since there is no supported API for constructing a trampoline class using the manual approach (sincedetail::nb_inst_python_derived()
is internal).