The new mechanism can deal with more complex scenarios such as
not nil field appearing in a non-default case object branch or
a field within a generic object that may depend on a when branch.
The commit also plugs another hole: the user is no longer able
to create illegal default values through seq.setLen(N).
* mark user defined destructors with sfOverriden to simplify the logic
* refactoring in preparation to merge liftings and generic instantiations for destructors
* ast: introduce nkHiddenTryStmt for destructor generation in order to be able to optimize better the code later on
* renamed 'patterns' switch to 'trmacros' as it was totally misleading before
* destructors: introduce tfCheckedForDestructor flag in preparation of strict =destroy checking
* test for invalid/too late destructor introductions
* liftdestructors: make code robust for nimsuggest
* --newruntime works for hello world again
* newruntime: code generation for closures
* kochdocs.nim: code cleanup
* docgen: nicer indentation
* parser.nim: code cleanup
* fixes#10458
* make tests green again
* make =destroy mixins
* gc:destructors: produced C code is almost working
* --gc:destructors simple program compiles (but leaks memory)
* gc:destructors make examples compile in C++ mode
* destructors: string implementation bugfixes
* strs.nim: minor code cleanup
* destructors: builtin seqs are beginning to work
* remove debugging helpers
Long story short, even if the type contains no reference at all to its
generic parameters we still have to walk its AST and evaluate any
nkRecWhen nodes that semRecordNodeAux skipped due to the type being a
generic one.
We also must be careful to modify the type `n` node in place since it
may be referenced by the caller as seen in the tillegaltyperecursion
test.
Moreover we also can't have the nkSym drift away from their original
values in order not to break the JS nkObjConstr codegen.
* late instantiation for the generic procs' default param values
* automatic mixin behaviour in concepts
Other fixes:
* don't render the automatically inserted default params in calls
* better rendering of tyFromExpr
* fix the usage of unresolved static[T] parameters in proc signatures
* fix tsametype and tmacrogenerics
* Allow creating composite type classes with concepts and using them in type signatures
* Allow integers to be used in ident concatenations
* Support using imported C++ generic types in proc signatures
* fixes#7230
* closes#7379
* re-enable some metatype tests
