essential for 1.0
=================

- introduce newSeqOfCap(10)
- The remaining bugs of the lambda lifting pass that is responsible to enable
  closures and closure iterators need to be fixed.
- ``concept`` needs to be refined, a nice name for the feature is not enough.
- Destructors need to be refined.
- make '--implicitStatic:on' the default; then we can also clean up the
  'static[T]' mess in the compiler!

- add "all threads are blocked" detection to 'spawn'
- Deprecate ``immediate`` for templates and macros
- document NimMain and check whether it works for threading
- remove   echo $foo  gotcha
- ``not`` or ``~`` for the effects system


Not critical for 1.0
====================

- figure out why C++ bootstrapping is so much slower
- The bitwise 'not' operator cold be renamed to 'bnot' to
  prevent 'not 4 == 5' from compiling. -> requires 'mixin' annotation for procs!
- make 'nil' work for 'add':
  - resizeString
  - incrSeq
  - addChar
- pragmas need 'bindSym' support
- allow simple read accesses to global variables --> difficult to ensure that
  no data races happen
- pragmas need re-work: 'push' is dangerous, 'hasPragma' does not work
  reliably with user-defined pragmas
- memory manager: add a measure of fragmentation
- we need a magic thisModule symbol
- optimize 'genericReset'; 'newException' leads to code bloat

- prevent 'alloc(TypeWithGCedMemory)'?
- split docgen into separate tool; not happy how it worked with ``nimsuggest``
  though
- map ``string`` and ``seq`` to ``std::string`` and ``std::vector``
- macro support for '='; bind '=' to a memory region
- macros as type pragmas


Bugs
====

- VM: Pegs do not work at compile-time
- VM: ptr/ref T cannot work in general
- blocks can "export" an identifier but the CCG generates {} for them ...
- ConcreteTypes in a 'case' means we don't check for duplicated case branches
- BUG: echo with template `$`*(info: TLineInfo): expr = toFileLineCol(info)


GC
==

- hybrid GC
- use big blocks in the allocator
- provide tool/API to track leaks/object counts
- resizing of strings/sequences could take into account the memory that
  is allocated


Concurrency
===========

- test 'deepCopy' for closures
- implement 'foo[1..4] = spawn(f[4..7])'

Low priority:
- support for exception propagation? (hard to implement)
- the copying of the 'ref Promise' into the thead local storage only
  happens to work due to the write barrier's implementation


CGEN
====
- codegen should use "NIM_CAST" macro and respect aliasing rules for GCC
- ``restrict`` pragma + backend support