* fixes#12015 by also checking kind of `typeNode`
If a tuple field is aliased it'll appear the same as a ref type in a
call to `getType` if only for the kind of the resulting `NimNode` is
checked (that is a `nnkBracketExpr`)
* fix test case due to #12017 and add more realistic test case
Adds an additional test case, which includes generics and is closer to
the real failure I encountered
* remove previous fix and fix differently after all
The previous fix was incomplete, because it failed for generics.
Note that the `of "tuple"` is not actually needed, the
`nnkBracketExpr` branch in the `else` branch would catch it too, but I
decided to introduce it for clarity. However, the latter is actually
needed, because it seems for aliases of `seq` we end up in it.
* update comment about global `%` proc in json test
* new .gensym implementation
* make astspec test green again
* introduce a --useVersion switch to group compatibility switches
* fixes#10180
* fixes#11494
* fixes#11483
* object constructor fields and named parameters are also not gensym'ed
* disabled broken package
* Add patch by @xenogenesi
* Async test for HTTP/1.1 without Content-Length
* Apply suggestions from code review
Co-Authored-By: Dominik Picheta <dominikpicheta@googlemail.com>
* IC: C codegen is aware of IC
* manual: minor change to make VSCode's RST plugin render it properly
* IC: minor refactoring
* testament: code refactorings
* rodutils: removed dead code
* IC: always build the compiler with the IC feature
* IC: C codegen improvements
* IC: implement the undocumented -d:nimMustCache option for testing purposes
* IC: added first basic tests
* IC: extensive testing of the deserialization feature
* testament: refactoring; better IC tests
* IC: removes 'nimMustCache' flag; readonly does the same
* testament: minor refactoring
* update Nimble version
* testament: removed dead code and imports; IC: added simple test
* IC: progress
