The type matching is done on the `T` of the `varargs[T]` so the
conversion must be performed to `T` and not to the whole type.
This problem is only noticeable with the cpp backend since C doesn't
give a damn shit about your fucking (wrong) types.
Fixes#9845
* compiler/parser: preserve lineinfo for accent quoted symbols
Previously the lineinfo for symbol $$$ in this example is:
proc `$$$`
^
After this commit:
proc `$$$`
^
* compiler/semstmts: correct lineinfo for accent quoted idents
Previously nimsuggest would highlight this as:
proc `$$$`
^~~
After this commit:
proc `$$$`
^~~
* nimsuggest/tests: add a test for accent quoted proc
Disabled by default
* Codegen now ignores object fields of type void
* Fix `$` bug for objects/tuples where it does not add a comma
* fields/fieldPairs iterators now ignore void types
* Use `isEmptyType` instead of checking for `tyVoid` directly
* fixes#8794 : `Error: undeclared field: 'foo'` should show type (+ where type is defined) (hard to guess in generic code)
* fixes#9270: `--listFullPaths` not honored by `declared in foo.nim` messages
* fixes#9767: VM stacktrace doesn't honor --excessiveStackTrace:on
* fixes#9768: VM stacktrace misses column info, can lead to ambiguous or harder to read stacktraces
* refactors some col+1 code to col + ColOffset (self documents code)
* make getProcHeader show declared info location also for types and all routine kinds (including macros,templates) instead of just (rather arbitrarily) for iterator,proc,func,method
* --listFullPaths now is honored in more places
* fix typo system/except.nim => lib/system/excpt.nim
* remove substr(foo, 0) hack in compiler/vm.nim which seems old and not applicable anymore
* add custom pragma support for var and let symbols
* updated changelog for custom pragmas on var and let symbols
* add oldast switch for backwards compatibility
* removed from `compiler`:
* lists (deprecated 2 years ago)
* removed from `lib` (all deprecated 3 years ago):
* ssl
* matchers
* httpserver
* removed from `lib/deprecated`:
* unsigned
* actors (and three accompanying tests)
* parseurl
* moved to `lib/deprecated`:
* securehash (the reason for not directly removing - it was deprecated (only) one year ago)
