path: root/todo.txt

version 0.9.2
=============

- fix destructors; don't work yet when used as expression; alternative for 
  version 1: disallow expressions yielding a type with a destructor that are
  not in a 'let/var' context  (p(a.openFile, b.openFile) makes no sense anyway)
- lazy overloading resolution:
  * get rid of ``expr[typ]``, use perhaps ``static[typ]`` instead
  * special case ``tyStmt``
- FFI:
  * test libffi on windows
  * test: times.format with the FFI
- acyclic vs prunable; introduce GC hints
- CGEN: ``restrict`` pragma + backend support; computed goto support
- document NimMain and check whether it works for threading
- a project wide override option for 'dynlib'


Bugs
====

- docgen: sometimes effects are listed twice
- even the easiest range checking is not performed anymore
- 'result' is not properly cleaned for NRVO
- instantiated generics are listed in error messages
- sneaking with qualifiedLookup() is really broken!
- aporia.nim(968, 5) Error: ambiguous identifier: 'DELETE' -- 
  use a qualifier
- blocks can "export" an identifier but the CCG generates {} for them ...
- osproc execProcesses can deadlock if all processes fail (as experienced
  in c++ mode)
- bootstrapping does not work in C++ mode


version 0.9.4
=============

- provide tool/API to track leaks/object counts
- hybrid GC
- use big blocks in the allocator
- implement full 'not nil' checking
- make 'bind' default for templates and introduce 'mixin';
  special rule for ``[]=``
- implicit deref for parameter matching; overloading based on 'var T'
- ``=`` should be overloadable; requires specialization for ``=``; general
  lift mechanism in the compiler is already implemented for 'fields'


version 0.9.X
=============

- test&finish first class iterators:
  * nested iterators
- implement the missing features wrt inheritance
- object pooling support for *hard* realtime systems
- improve the compiler as a service
- better support for macros that rewrite procs
- macros need access to types and symbols (partially implemented)
- result = result shr 8 for the "system()" wrapper
- rethink the syntax/grammar:
  * parser is not strict enough with newlines
  * change comment handling in the AST


Concurrency
-----------

- shared memory heap: ``shared ref`` etc. The only hard part in the GC is to
  "stop the world". However, it may be worthwhile to generate explicit 
  (or implicit) syncGC() calls in loops. Automatic loop injection seems
  troublesome, but maybe we can come up with a simple heuristic. (All procs
  that `new` shared memory are syncGC() candidates... But then 'new' itself
  calls syncGC() so that's pointless.) Hm instead of an heuristic simply
  provide a ``syncgc`` pragma to trigger compiler injection --> more general:
  an ``injectLoop`` pragma
- 'writes: []' effect; track reads/writes for shared types
- use the effect system for static deadlock prevention and race detection
- ``~`` operator for effects
- introduce 'noaddr' pragma to prevent taking the address of a location; this
  is very handy to prevent aliasing of global data


version 0.9.XX
==============

- object branch transitions can't work with the current 'reset'; add a 'reset'
  with an additional parameter --> simple:
  provide a 'reset(x, TObj(k: nkValue))' instead? why bother? '=' does the
  same.
- document nimdoc properly finally
- make 'clamp' a magic for the range stuff
- better type syntax for functions and tuples: tuple(int, int); (int,int)->int


Not essential for 1.0.0
=======================

- 'const' objects including case objects
- mocking support with ``tyProxy`` that does: fallback for ``.`` operator
  - overloading of ``.``? Special case ``.=``?
- allow implicit forward declarations of procs via a pragma (so that the
  wrappers can deactivate it)
- implement the "snoopResult" pragma; no, make a strutils with string append
  semantics instead ...
- implement "closure tuple consists of a single 'ref'" optimization
- optimize method dispatchers
- ``with proc `+`(x, y: T): T`` for generic code
- new feature: ``distinct T with operations``
- arglist as a type (iterator chaining); variable length type lists for generics
- resizing of strings/sequences could take into account the memory that
  is allocated
- codegen should use "NIM_CAST" macro and respect aliasing rules for GCC
- implement marker procs for message passing
- activate more thread tests
- implement closures that support nesting of *procs* > 1
- GC: implement simple generational GC
  * first version: mark black in write barrier
  * second version: introduce fake roots instead of marking black
  * third version: find some clever heuristic which is preferable
- object constructors: static check for fields if discriminator is known at 
  compile time


GC
==

- precise stack marking; embrace C++ code generation for that
- marker procs for Boehm GC
- implement 'mixed' GC mode


Optimizations
=============

- optimize 'if' with a constant condition --> necessary in frontend for better
  dead code elimination
- escape analysis for string/seq seems to be easy to do too;
  even further write barrier specialization
- inlining of first class functions
- proc specialization in the code gen for write barrier specialization