changelog_2_2_0.md 9.0 KB

v2.2.0 - 2024-10-02

Changes affecting backward compatibility

  • -d:nimStrictDelete becomes the default. An index error is produced when the index passed to system.delete is out of bounds. Use -d:nimAuditDelete to mimic the old behavior for backward compatibility.

  • The default user-agent in std/httpclient has been changed to Nim-httpclient/<version> instead of Nim httpclient/<version> which was incorrect according to the HTTP spec.

  • Methods now support implementations based on a VTable by using --experimental:vtables. Methods are then confined to the same module where their type has been defined.

  • With -d:nimPreviewNonVarDestructor, non-var destructors become the default.

  • A bug where tuple unpacking assignment with a longer tuple on the RHS than the LHS was allowed has been fixed, i.e. code like:

    var a, b: int
    (a, b) = (1, 2, 3, 4)
    

    will no longer compile.

  • internalNew is removed from the system module, use new instead.

  • bindMethod in std/jsffi is deprecated, don't use it with closures.

  • The JS backend now supports lambda lifting for closures. Use --legacy:jsNoLambdaLifting to emulate old behaviors.

  • The JS backend now supports closure iterators.

  • owner in std/macros is deprecated.

  • Ambiguous type symbols in generic procs and templates now generate symchoice nodes. Previously; in templates they would error immediately at the template definition, and in generic procs a type symbol would arbitrarily be captured, losing the information of the other symbols. This means that generic code can now give errors for ambiguous type symbols, and macros operating on generic proc AST may encounter symchoice nodes instead of the arbitrarily resolved type symbol nodes.

  • Partial generic instantiation of routines is no longer allowed. Previously it compiled in niche situations due to bugs in the compiler.

  proc foo[T, U](x: T, y: U) = echo (x, y)
  proc foo[T, U](x: var T, y: U) = echo "var ", (x, y)

  proc bar[T]() =
    foo[float](1, "abc")

  bar[int]() # before: (1.0, "abc"), now: type mismatch, missing generic parameter
  • const values now open a new scope for each constant, meaning symbols declared in them can no longer be used outside or in the value of other constants.
  const foo = (var a = 1; a)
  const bar = a # error
  let baz = a # error
  • The following POSIX wrappers have had their types changed from signed to unsigned types on OSX and FreeBSD/OpenBSD to correct codegen errors:

    • Gid (was int32, is now uint32)
    • Uid (was int32, is now uint32)
    • Dev (was int32, is now uint32 on FreeBSD)
    • Nlink (was int16, is now uint32 on OpenBSD and uint16 on OSX/other BSD)
    • sin6_flowinfo and sin6_scope_id fields of Sockaddr_in6 (were int32, are now uint32)
    • n_net field of Tnetent (was int32, is now uint32)
  • The Atomic[T] type on C++ now uses C11 primitives by default instead of std::atomic. To use std::atomic instead, -d:nimUseCppAtomics can be defined.

Standard library additions and changes

  • Changed std/osfiles.copyFile to allow specifying bufferSize instead of a hard-coded one.
  • Changed std/osfiles.copyFile to use POSIX_FADV_SEQUENTIAL hints for kernel-level aggressive sequential read-aheads.
  • std/htmlparser has been moved to a nimble package, use nimble or atlas to install it.
  • Changed std/os.copyDir and copyDirWithPermissions to allow skipping special "file" objects like FIFOs, device files, etc on Unix by specifying a skipSpecial parameter.

  • Added newStringUninit to the system module, which creates a new string of length len like newString but with uninitialized content.

  • Added setLenUninit to the system module, which doesn't initialize slots when enlarging a sequence.

  • Added hasDefaultValue to std/typetraits to check if a type has a valid default value.

  • Added rangeBase to std/typetraits to obtain the base type of a range type or convert a value with a range type to its base type.

  • Added Viewport API for the JavaScript targets in the dom module.

  • Added toSinglyLinkedRing and toDoublyLinkedRing to std/lists to convert from openArrays.

  • ORC: To be enabled via nimOrcStats there is a new API called GC_orcStats that can be used to query how many objects the cyclic collector did free. If the number is zero that is a strong indicator that you can use --mm:arc instead of --mm:orc.

  • A $ template is provided for Path in std/paths.

  • std/hashes.hash(x:string) changed to produce a 64-bit string Hash (based on Google's Farm Hash) which is also often faster than the present one. Define nimStringHash2 to get the old values back. --jsbigint=off mode always only produces the old values. This may impact your automated tests if they depend on hash order in some obvious or indirect way. Using sorted or OrderedTable is often an easy workaround.

  • Deprecates system.newSeqUninitialized, which is replaced by newSeqUninit.

Language changes

  • noInit can be used in types and fields to disable member initializers in the C++ backend.

  • C++ custom constructors initializers see https://nim-lang.org/docs/manual_experimental.html#constructor-initializer

  • member can be used to attach a procedure to a C++ type.

  • Inside a C++ constructor, result can be used to access the created object rather than this.

  • Tuple unpacking changes:

    • Tuple unpacking assignment now supports using underscores to discard values. nim var a, c: int (a, _, c) = (1, 2, 3)
    • Tuple unpacking variable declarations now support type annotations, but only for the entire tuple. nim let (a, b): (int, int) = (1, 2) let (a, (b, c)): (byte, (float, cstring)) = (1, (2, "abc"))
  • The experimental option --experimental:openSym has been added to allow captured symbols in generic routine and template bodies respectively to be replaced by symbols injected locally by templates/macros at instantiation time. bind may be used to keep the captured symbols over the injected ones regardless of enabling the option, but other methods like renaming the captured symbols should be used instead so that the code is not affected by context changes.

Since this change may affect runtime behavior, the experimental switch openSym needs to be enabled; and a warning is given in the case where an injected symbol would replace a captured symbol not bound by bind and the experimental switch isn't enabled.

  const value = "captured"
  template foo(x: int, body: untyped): untyped =
    let value {.inject.} = "injected"
    body

  proc old[T](): string =
    foo(123):
      return value # warning: a new `value` has been injected, use `bind` or turn on `experimental:openSym`
  echo old[int]() # "captured"

  template oldTempl(): string =
    block:
      foo(123):
        value # warning: a new `value` has been injected, use `bind` or turn on `experimental:openSym`
  echo oldTempl() # "captured"

  {.experimental: "openSym".}

  proc bar[T](): string =
    foo(123):
      return value
  assert bar[int]() == "injected" # previously it would be "captured"

  proc baz[T](): string =
    bind value
    foo(123):
      return value
  assert baz[int]() == "captured"

  template barTempl(): string =
    block:
      foo(123):
        value
  assert barTempl() == "injected" # previously it would be "captured"

  template bazTempl(): string =
    bind value
    block:
      foo(123):
        value
  assert bazTempl() == "captured"

This option also generates a new node kind nnkOpenSym which contains exactly 1 nnkSym node. In the future this might be merged with a slightly modified nnkOpenSymChoice node but macros that want to support the experimental feature should still handle nnkOpenSym, as the node kind would simply not be generated as opposed to being removed.

Another experimental switch genericsOpenSym exists that enables this behavior at instantiation time, meaning templates etc can enable it specifically when they are being called. However this does not generate nnkOpenSym nodes (unless the other switch is enabled) and so doesn't reflect the regular behavior of the switch.

  const value = "captured"
  template foo(x: int, body: untyped): untyped =
    let value {.inject.} = "injected"
    {.push experimental: "genericsOpenSym".}
    body
    {.pop.}

  proc bar[T](): string =
    foo(123):
      return value
  echo bar[int]() # "injected"

  template barTempl(): string =
    block:
      var res: string
      foo(123):
        res = value
      res
  assert barTempl() == "injected"

Compiler changes

  • --nimcache using a relative path as the argument in a config file is now relative to the config file instead of the current directory.

Tool changes

  • koch now allows bootstrapping with -d:nimHasLibFFI, replacing the older option of building the compiler directly w/ the libffi nimble package.