diff options
Diffstat (limited to 'lib/core')
| -rw-r--r-- | lib/core/hotcodereloading.nim | 2 | ||||
| -rw-r--r-- | lib/core/locks.nim | 28 | ||||
| -rw-r--r-- | lib/core/macrocache.nim | 230 | ||||
| -rw-r--r-- | lib/core/macros.nim | 1085 | ||||
| -rw-r--r-- | lib/core/rlocks.nim | 28 | ||||
| -rw-r--r-- | lib/core/typeinfo.nim | 352 |
6 files changed, 1019 insertions, 706 deletions
diff --git a/lib/core/hotcodereloading.nim b/lib/core/hotcodereloading.nim index 73f38402d..3a876885c 100644 --- a/lib/core/hotcodereloading.nim +++ b/lib/core/hotcodereloading.nim @@ -11,7 +11,7 @@ when defined(hotcodereloading): import - macros + std/macros template beforeCodeReload*(body: untyped) = hcrAddEventHandler(true, proc = body) {.executeOnReload.} diff --git a/lib/core/locks.nim b/lib/core/locks.nim index bddd6d864..523727479 100644 --- a/lib/core/locks.nim +++ b/lib/core/locks.nim @@ -18,8 +18,7 @@ when not compileOption("threads") and not defined(nimdoc): when false: # fix #12330 {.error: "Locks requires --threads:on option.".} -const insideRLocksModule = false -include "system/syslocks" +import std/private/syslocks type Lock* = SysLock ## Nim lock; whether this is re-entrant @@ -28,25 +27,30 @@ type {.push stackTrace: off.} + +proc `$`*(lock: Lock): string = + # workaround bug #14873 + result = "()" + proc initLock*(lock: var Lock) {.inline.} = ## Initializes the given lock. when not defined(js): initSysLock(lock) -proc deinitLock*(lock: var Lock) {.inline.} = +proc deinitLock*(lock: Lock) {.inline.} = ## Frees the resources associated with the lock. deinitSys(lock) -proc tryAcquire*(lock: var Lock): bool = +proc tryAcquire*(lock: var Lock): bool {.inline.} = ## Tries to acquire the given lock. Returns `true` on success. result = tryAcquireSys(lock) -proc acquire*(lock: var Lock) = +proc acquire*(lock: var Lock) {.inline.} = ## Acquires the given lock. when not defined(js): acquireSys(lock) -proc release*(lock: var Lock) = +proc release*(lock: var Lock) {.inline.} = ## Releases the given lock. when not defined(js): releaseSys(lock) @@ -56,22 +60,26 @@ proc initCond*(cond: var Cond) {.inline.} = ## Initializes the given condition variable. initSysCond(cond) -proc deinitCond*(cond: var Cond) {.inline.} = +proc deinitCond*(cond: Cond) {.inline.} = ## Frees the resources associated with the condition variable. deinitSysCond(cond) proc wait*(cond: var Cond, lock: var Lock) {.inline.} = - ## waits on the condition variable `cond`. + ## Waits on the condition variable `cond`. waitSysCond(cond, lock) proc signal*(cond: var Cond) {.inline.} = - ## sends a signal to the condition variable `cond`. + ## Sends a signal to the condition variable `cond`. signalSysCond(cond) +proc broadcast*(cond: var Cond) {.inline.} = + ## Unblocks all threads currently blocked on the + ## specified condition variable `cond`. + broadcastSysCond(cond) + template withLock*(a: Lock, body: untyped) = ## Acquires the given lock, executes the statements in body and ## releases the lock after the statements finish executing. - mixin acquire, release acquire(a) {.locks: [a].}: try: diff --git a/lib/core/macrocache.nim b/lib/core/macrocache.nim index 8fe1fa603..39999fa11 100644 --- a/lib/core/macrocache.nim +++ b/lib/core/macrocache.nim @@ -7,38 +7,238 @@ # distribution, for details about the copyright. # -## This module provides an API for macros that need to collect compile -## time information across module boundaries in global variables. -## Starting with version 0.19 of Nim this is not directly supported anymore -## as it breaks incremental compilations. -## Instead the API here needs to be used. +## This module provides an API for macros to collect compile-time information +## across module boundaries. It should be used instead of global `{.compileTime.}` +## variables as those break incremental compilation. +## +## The main feature of this module is that if you create `CacheTable`s or +## any other `Cache` types with the same name in different modules, their +## content will be shared, meaning that you can fill a `CacheTable` in +## one module, and iterate over its contents in another. + +runnableExamples: + import std/macros + + const mcTable = CacheTable"myTable" + const mcSeq = CacheSeq"mySeq" + const mcCounter = CacheCounter"myCounter" + + static: + # add new key "val" with the value `myval` + let myval = newLit("hello ic") + mcTable["val"] = myval + assert mcTable["val"].kind == nnkStrLit + + # Can access the same cache from different static contexts + # All the information is retained + static: + # get value from `mcTable` and add it to `mcSeq` + mcSeq.add(mcTable["val"]) + assert mcSeq.len == 1 + + static: + assert mcSeq[0].strVal == "hello ic" + + # increase `mcCounter` by 3 + mcCounter.inc(3) + assert mcCounter.value == 3 + type CacheSeq* = distinct string + ## Compile-time sequence of `NimNode`s. CacheTable* = distinct string + ## Compile-time table of key-value pairs. + ## + ## Keys are `string`s and values are `NimNode`s. CacheCounter* = distinct string + ## Compile-time counter, uses `int` for storing the count. + +proc value*(c: CacheCounter): int {.magic: "NccValue".} = + ## Returns the value of a counter `c`. + runnableExamples: + static: + let counter = CacheCounter"valTest" + # default value is 0 + assert counter.value == 0 + + inc counter + assert counter.value == 1 + +proc inc*(c: CacheCounter; by = 1) {.magic: "NccInc".} = + ## Increments the counter `c` with the value `by`. + runnableExamples: + static: + let counter = CacheCounter"incTest" + inc counter + inc counter, 5 + + assert counter.value == 6 + +proc add*(s: CacheSeq; value: NimNode) {.magic: "NcsAdd".} = + ## Adds `value` to `s`. + runnableExamples: + import std/macros + const mySeq = CacheSeq"addTest" + + static: + mySeq.add(newLit(5)) + mySeq.add(newLit("hello ic")) + + assert mySeq.len == 2 + assert mySeq[1].strVal == "hello ic" + +proc incl*(s: CacheSeq; value: NimNode) {.magic: "NcsIncl".} = + ## Adds `value` to `s`. + ## + ## .. hint:: This doesn't do anything if `value` is already in `s`. + runnableExamples: + import std/macros + const mySeq = CacheSeq"inclTest" + + static: + mySeq.incl(newLit(5)) + mySeq.incl(newLit(5)) + + # still one element + assert mySeq.len == 1 + +proc len*(s: CacheSeq): int {.magic: "NcsLen".} = + ## Returns the length of `s`. + runnableExamples: + import std/macros -proc value*(c: CacheCounter): int {.magic: "NccValue".} -proc inc*(c: CacheCounter; by = 1) {.magic: "NccInc".} + const mySeq = CacheSeq"lenTest" + static: + let val = newLit("helper") + mySeq.add(val) + assert mySeq.len == 1 -proc add*(s: CacheSeq; value: NimNode) {.magic: "NcsAdd".} -proc incl*(s: CacheSeq; value: NimNode) {.magic: "NcsIncl".} -proc len*(s: CacheSeq): int {.magic: "NcsLen".} -proc `[]`*(s: CacheSeq; i: int): NimNode {.magic: "NcsAt".} + mySeq.add(val) + assert mySeq.len == 2 + +proc `[]`*(s: CacheSeq; i: int): NimNode {.magic: "NcsAt".} = + ## Returns the `i`th value from `s`. + runnableExamples: + import std/macros + + const mySeq = CacheSeq"subTest" + static: + mySeq.add(newLit(42)) + assert mySeq[0].intVal == 42 + +proc `[]`*(s: CacheSeq; i: BackwardsIndex): NimNode = + ## Returns the `i`th last value from `s`. + runnableExamples: + import std/macros + + const mySeq = CacheSeq"backTest" + static: + mySeq &= newLit(42) + mySeq &= newLit(7) + assert mySeq[^1].intVal == 7 # Last item + assert mySeq[^2].intVal == 42 # Second last item + s[s.len - int(i)] iterator items*(s: CacheSeq): NimNode = + ## Iterates over each item in `s`. + runnableExamples: + import std/macros + const myseq = CacheSeq"itemsTest" + + static: + myseq.add(newLit(5)) + myseq.add(newLit(42)) + + for val in myseq: + # check that all values in `myseq` are int literals + assert val.kind == nnkIntLit + for i in 0 ..< len(s): yield s[i] -proc `[]=`*(t: CacheTable; key: string, value: NimNode) {.magic: "NctPut".} - ## 'key' has to be unique! +proc `[]=`*(t: CacheTable; key: string, value: NimNode) {.magic: "NctPut".} = + ## Inserts a `(key, value)` pair into `t`. + ## + ## .. warning:: `key` has to be unique! Assigning `value` to a `key` that is already + ## in the table will result in a compiler error. + runnableExamples: + import std/macros + + const mcTable = CacheTable"subTest" + static: + # assign newLit(5) to the key "value" + mcTable["value"] = newLit(5) + + # check that we can get the value back + assert mcTable["value"].kind == nnkIntLit -proc len*(t: CacheTable): int {.magic: "NctLen".} -proc `[]`*(t: CacheTable; key: string): NimNode {.magic: "NctGet".} +proc len*(t: CacheTable): int {.magic: "NctLen".} = + ## Returns the number of elements in `t`. + runnableExamples: + import std/macros + + const dataTable = CacheTable"lenTest" + static: + dataTable["key"] = newLit(5) + assert dataTable.len == 1 + +proc `[]`*(t: CacheTable; key: string): NimNode {.magic: "NctGet".} = + ## Retrieves the `NimNode` value at `t[key]`. + runnableExamples: + import std/macros + + const mcTable = CacheTable"subTest" + static: + mcTable["toAdd"] = newStmtList() + + # get the NimNode back + assert mcTable["toAdd"].kind == nnkStmtList + +proc hasKey*(t: CacheTable; key: string): bool = + ## Returns true if `key` is in the table `t`. + ## + ## See also: + ## * [contains proc][contains(CacheTable, string)] for use with the `in` operator + runnableExamples: + import std/macros + const mcTable = CacheTable"hasKeyEx" + static: + assert not mcTable.hasKey("foo") + mcTable["foo"] = newEmptyNode() + # Will now be true since we inserted a value + assert mcTable.hasKey("foo") + discard "Implemented in vmops" + +proc contains*(t: CacheTable; key: string): bool {.inline.} = + ## Alias of [hasKey][hasKey(CacheTable, string)] for use with the `in` operator. + runnableExamples: + import std/macros + const mcTable = CacheTable"containsEx" + static: + mcTable["foo"] = newEmptyNode() + # Will be true since we gave it a value before + assert "foo" in mcTable + t.hasKey(key) proc hasNext(t: CacheTable; iter: int): bool {.magic: "NctHasNext".} proc next(t: CacheTable; iter: int): (string, NimNode, int) {.magic: "NctNext".} iterator pairs*(t: CacheTable): (string, NimNode) = + ## Iterates over all `(key, value)` pairs in `t`. + runnableExamples: + import std/macros + const mytabl = CacheTable"values" + + static: + mytabl["intVal"] = newLit(5) + mytabl["otherVal"] = newLit(6) + for key, val in mytabl: + # make sure that we actually get the same keys + assert key in ["intVal", "otherVal"] + + # all vals are int literals + assert val.kind == nnkIntLit + var h = 0 while hasNext(t, h): let (a, b, h2) = next(t, h) diff --git a/lib/core/macros.nim b/lib/core/macros.nim index 97c3a46c5..7646b165c 100644 --- a/lib/core/macros.nim +++ b/lib/core/macros.nim @@ -10,6 +10,10 @@ include "system/inclrtl" import std/private/since +when defined(nimPreviewSlimSystem): + import std/[assertions, formatfloat] + + ## This module contains the interface to the compiler's abstract syntax ## tree (`AST`:idx:). Macros operate on this tree. ## @@ -19,6 +23,8 @@ import std/private/since ## .. include:: ../../doc/astspec.txt +## .. importdoc:: system.nim + # If you look for the implementation of the magic symbol # ``{.magic: "Foo".}``, search for `mFoo` and `opcFoo`. @@ -71,21 +77,24 @@ type nnkTupleTy, nnkTupleClassTy, nnkTypeClassTy, nnkStaticTy, nnkRecList, nnkRecCase, nnkRecWhen, nnkRefTy, nnkPtrTy, nnkVarTy, - nnkConstTy, nnkMutableTy, + nnkConstTy, nnkOutTy, nnkDistinctTy, nnkProcTy, nnkIteratorTy, # iterator type - nnkSharedTy, # 'shared T' + nnkSinkAsgn, nnkEnumTy, nnkEnumFieldDef, - nnkArglist, nnkPattern + nnkArgList, nnkPattern nnkHiddenTryStmt, nnkClosure, nnkGotoState, nnkState, nnkBreakState, nnkFuncDef, - nnkTupleConstr + nnkTupleConstr, + nnkError, ## erroneous AST node + nnkModuleRef, nnkReplayAction, nnkNilRodNode ## internal IC nodes + nnkOpenSym NimNodeKinds* = set[NimNodeKind] NimTypeKind* = enum # some types are no longer used, see ast.nim @@ -120,11 +129,15 @@ type TNimSymKinds* {.deprecated.} = set[NimSymKind] +const + nnkMutableTy* {.deprecated.} = nnkOutTy + nnkSharedTy* {.deprecated.} = nnkSinkAsgn + type NimIdent* {.deprecated.} = object of RootObj ## Represents a Nim identifier in the AST. **Note**: This is only ## rarely useful, for identifier construction from a string - ## use ``ident"abc"``. + ## use `ident"abc"`. NimSymObj = object # hidden NimSym* {.deprecated.} = ref NimSymObj @@ -134,8 +147,9 @@ type const nnkLiterals* = {nnkCharLit..nnkNilLit} + # see matching set CallNodes below nnkCallKinds* = {nnkCall, nnkInfix, nnkPrefix, nnkPostfix, nnkCommand, - nnkCallStrLit} + nnkCallStrLit, nnkHiddenCallConv} nnkPragmaCallKinds = {nnkExprColonExpr, nnkCall, nnkCallStrLit} {.push warnings: off.} @@ -175,9 +189,9 @@ proc `[]`*(n: NimNode, i: BackwardsIndex): NimNode = n[n.len - i.int] template `^^`(n: NimNode, i: untyped): untyped = (when i is BackwardsIndex: n.len - int(i) else: int(i)) -proc `[]`*[T, U](n: NimNode, x: HSlice[T, U]): seq[NimNode] = +proc `[]`*[T, U: Ordinal](n: NimNode, x: HSlice[T, U]): seq[NimNode] = ## Slice operation for NimNode. - ## Returns a seq of child of `n` who inclusive range [n[x.a], n[x.b]]. + ## Returns a seq of child of `n` who inclusive range `[n[x.a], n[x.b]]`. let xa = n ^^ x.a let L = (n ^^ x.b) - xa + 1 result = newSeq[NimNode](L) @@ -193,13 +207,12 @@ proc `[]=`*(n: NimNode, i: BackwardsIndex, child: NimNode) = n[n.len - i.int] = child template `or`*(x, y: NimNode): NimNode = - ## Evaluate ``x`` and when it is not an empty node, return - ## it. Otherwise evaluate to ``y``. Can be used to chain several + ## Evaluate `x` and when it is not an empty node, return + ## it. Otherwise evaluate to `y`. Can be used to chain several ## expressions to get the first expression that is not empty. - ## - ## .. code-block:: nim - ## + ## ```nim ## let node = mightBeEmpty() or mightAlsoBeEmpty() or fallbackNode + ## ``` let arg = x if arg != nil and arg.kind != nnkEmpty: @@ -208,12 +221,12 @@ template `or`*(x, y: NimNode): NimNode = y proc add*(father, child: NimNode): NimNode {.magic: "NAdd", discardable, - noSideEffect, locks: 0.} + noSideEffect.} ## Adds the `child` to the `father` node. Returns the ## father node so that calls can be nested. proc add*(father: NimNode, children: varargs[NimNode]): NimNode {. - magic: "NAddMultiple", discardable, noSideEffect, locks: 0.} + magic: "NAddMultiple", discardable, noSideEffect.} ## Adds each child of `children` to the `father` node. ## Returns the `father` node so that calls can be nested. @@ -229,7 +242,17 @@ proc intVal*(n: NimNode): BiggestInt {.magic: "NIntVal", noSideEffect.} proc floatVal*(n: NimNode): BiggestFloat {.magic: "NFloatVal", noSideEffect.} ## Returns a float from any floating point literal. -{.push warnings: off.} + +proc symKind*(symbol: NimNode): NimSymKind {.magic: "NSymKind", noSideEffect.} +proc getImpl*(symbol: NimNode): NimNode {.magic: "GetImpl", noSideEffect.} + ## Returns a copy of the declaration of a symbol or `nil`. +proc strVal*(n: NimNode): string {.magic: "NStrVal", noSideEffect.} + ## Returns the string value of an identifier, symbol, comment, or string literal. + ## + ## See also: + ## * `strVal= proc<#strVal=,NimNode,string>`_ for setting the string value. + +{.push warnings: off.} # silence `deprecated` proc ident*(n: NimNode): NimIdent {.magic: "NIdent", noSideEffect, deprecated: "Deprecated since version 0.18.1; All functionality is defined on 'NimNode'.".} @@ -239,41 +262,13 @@ proc symbol*(n: NimNode): NimSym {.magic: "NSymbol", noSideEffect, deprecated: proc getImpl*(s: NimSym): NimNode {.magic: "GetImpl", noSideEffect, deprecated: "use `getImpl: NimNode -> NimNode` instead".} -when defined(nimSymKind): - proc symKind*(symbol: NimNode): NimSymKind {.magic: "NSymKind", noSideEffect.} - proc getImpl*(symbol: NimNode): NimNode {.magic: "GetImpl", noSideEffect.} - ## Returns a copy of the declaration of a symbol or `nil`. - proc strVal*(n: NimNode): string {.magic: "NStrVal", noSideEffect.} - ## Returns the string value of an identifier, symbol, comment, or string literal. - ## - ## See also: - ## * `strVal= proc<#strVal=,NimNode,string>`_ for setting the string value. - - proc `$`*(i: NimIdent): string {.magic: "NStrVal", noSideEffect, deprecated: - "Deprecated since version 0.18.1; Use 'strVal' instead.".} - ## Converts a Nim identifier to a string. - - proc `$`*(s: NimSym): string {.magic: "NStrVal", noSideEffect, deprecated: - "Deprecated since version 0.18.1; Use 'strVal' instead.".} - ## Converts a Nim symbol to a string. +proc `$`*(i: NimIdent): string {.magic: "NStrVal", noSideEffect, deprecated: + "Deprecated since version 0.18.1; Use 'strVal' instead.".} + ## Converts a Nim identifier to a string. -else: # bootstrapping substitute - proc getImpl*(symbol: NimNode): NimNode = - symbol.symbol.getImpl - - proc strValOld(n: NimNode): string {.magic: "NStrVal", noSideEffect.} - - proc `$`*(s: NimSym): string {.magic: "IdentToStr", noSideEffect.} - - proc `$`*(i: NimIdent): string {.magic: "IdentToStr", noSideEffect.} - - proc strVal*(n: NimNode): string = - if n.kind == nnkIdent: - $n.ident - elif n.kind == nnkSym: - $n.symbol - else: - n.strValOld +proc `$`*(s: NimSym): string {.magic: "NStrVal", noSideEffect, deprecated: + "Deprecated since version 0.18.1; Use 'strVal' instead.".} + ## Converts a Nim symbol to a string. {.pop.} @@ -284,23 +279,21 @@ when (NimMajor, NimMinor, NimPatch) >= (1, 3, 5) or defined(nimSymImplTransform) ## note that code transformations are implementation dependent and subject to change. ## See an example in `tests/macros/tmacros_various.nim`. -when defined(nimHasSymOwnerInMacro): - proc owner*(sym: NimNode): NimNode {.magic: "SymOwner", noSideEffect.} - ## Accepts a node of kind `nnkSym` and returns its owner's symbol. - ## The meaning of 'owner' depends on `sym`'s `NimSymKind` and declaration - ## context. For top level declarations this is an `nskModule` symbol, - ## for proc local variables an `nskProc` symbol, for enum/object fields an - ## `nskType` symbol, etc. For symbols without an owner, `nil` is returned. - ## - ## See also: - ## * `symKind proc<#symKind,NimNode>`_ to get the kind of a symbol - ## * `getImpl proc<#getImpl,NimNode>`_ to get the declaration of a symbol - -when defined(nimHasInstantiationOfInMacro): - proc isInstantiationOf*(instanceProcSym, genProcSym: NimNode): bool {.magic: "SymIsInstantiationOf", noSideEffect.} - ## Checks if a proc symbol is an instance of the generic proc symbol. - ## Useful to check proc symbols against generic symbols - ## returned by `bindSym`. +proc owner*(sym: NimNode): NimNode {.magic: "SymOwner", noSideEffect, deprecated.} + ## Accepts a node of kind `nnkSym` and returns its owner's symbol. + ## The meaning of 'owner' depends on `sym`'s `NimSymKind` and declaration + ## context. For top level declarations this is an `nskModule` symbol, + ## for proc local variables an `nskProc` symbol, for enum/object fields an + ## `nskType` symbol, etc. For symbols without an owner, `nil` is returned. + ## + ## See also: + ## * `symKind proc<#symKind,NimNode>`_ to get the kind of a symbol + ## * `getImpl proc<#getImpl,NimNode>`_ to get the declaration of a symbol + +proc isInstantiationOf*(instanceProcSym, genProcSym: NimNode): bool {.magic: "SymIsInstantiationOf", noSideEffect.} + ## Checks if a proc symbol is an instance of the generic proc symbol. + ## Useful to check proc symbols against generic symbols + ## returned by `bindSym`. proc getType*(n: NimNode): NimNode {.magic: "NGetType", noSideEffect.} ## With 'getType' you can access the node's `type`:idx:. A Nim type is @@ -311,11 +304,11 @@ proc getType*(n: NimNode): NimNode {.magic: "NGetType", noSideEffect.} ## kind of type it is, call `typeKind` on getType's result. proc getType*(n: typedesc): NimNode {.magic: "NGetType", noSideEffect.} - ## Version of ``getType`` which takes a ``typedesc``. + ## Version of `getType` which takes a `typedesc`. proc typeKind*(n: NimNode): NimTypeKind {.magic: "NGetType", noSideEffect.} ## Returns the type kind of the node 'n' that should represent a type, that - ## means the node should have been obtained via ``getType``. + ## means the node should have been obtained via `getType`. proc getTypeInst*(n: NimNode): NimNode {.magic: "NGetType", noSideEffect.} = ## Returns the `type`:idx: of a node in a form matching the way the @@ -336,12 +329,12 @@ proc getTypeInst*(n: NimNode): NimNode {.magic: "NGetType", noSideEffect.} = doAssert(dumpTypeInst(c) == "Vec[4, float32]") proc getTypeInst*(n: typedesc): NimNode {.magic: "NGetType", noSideEffect.} - ## Version of ``getTypeInst`` which takes a ``typedesc``. + ## Version of `getTypeInst` which takes a `typedesc`. proc getTypeImpl*(n: NimNode): NimNode {.magic: "NGetType", noSideEffect.} = ## Returns the `type`:idx: of a node in a form matching the implementation ## of the type. Any intermediate aliases are expanded to arrive at the final - ## type implementation. You can instead use ``getImpl`` on a symbol if you + ## type implementation. You can instead use `getImpl` on a symbol if you ## want to find the intermediate aliases. runnableExamples: type @@ -356,18 +349,16 @@ proc getTypeImpl*(n: NimNode): NimNode {.magic: "NGetType", noSideEffect.} = newLit(x.getTypeImpl.repr) let t = """ object - arr: array[0 .. 3, float32] -""" + arr: array[0 .. 3, float32]""" doAssert(dumpTypeImpl(a) == t) doAssert(dumpTypeImpl(b) == t) doAssert(dumpTypeImpl(c) == t) -when defined(nimHasSignatureHashInMacro): - proc signatureHash*(n: NimNode): string {.magic: "NSigHash", noSideEffect.} - ## Returns a stable identifier derived from the signature of a symbol. - ## The signature combines many factors such as the type of the symbol, - ## the owning module of the symbol and others. The same identifier is - ## used in the back-end to produce the mangled symbol name. +proc signatureHash*(n: NimNode): string {.magic: "NSigHash", noSideEffect.} + ## Returns a stable identifier derived from the signature of a symbol. + ## The signature combines many factors such as the type of the symbol, + ## the owning module of the symbol and others. The same identifier is + ## used in the back-end to produce the mangled symbol name. proc symBodyHash*(s: NimNode): string {.noSideEffect.} = ## Returns a stable digest for symbols derived not only from type signature @@ -377,7 +368,7 @@ proc symBodyHash*(s: NimNode): string {.noSideEffect.} = discard proc getTypeImpl*(n: typedesc): NimNode {.magic: "NGetType", noSideEffect.} - ## Version of ``getTypeImpl`` which takes a ``typedesc``. + ## Version of `getTypeImpl` which takes a `typedesc`. proc `intVal=`*(n: NimNode, val: BiggestInt) {.magic: "NSetIntVal", noSideEffect.} proc `floatVal=`*(n: NimNode, val: BiggestFloat) {.magic: "NSetFloatVal", noSideEffect.} @@ -407,15 +398,43 @@ proc newNimNode*(kind: NimNodeKind, {.magic: "NNewNimNode", noSideEffect.} ## Creates a new AST node of the specified kind. ## - ## The ``lineInfoFrom`` parameter is used for line information when the + ## The `lineInfoFrom` parameter is used for line information when the ## produced code crashes. You should ensure that it is set to a node that ## you are transforming. -proc copyNimNode*(n: NimNode): NimNode {.magic: "NCopyNimNode", noSideEffect.} -proc copyNimTree*(n: NimNode): NimNode {.magic: "NCopyNimTree", noSideEffect.} +proc copyNimNode*(n: NimNode): NimNode {.magic: "NCopyNimNode", noSideEffect.} = + ## Creates a new AST node by copying the node `n`. Note that unlike `copyNimTree`, + ## child nodes of `n` are not copied. + runnableExamples: + macro foo(x: typed) = + var s = copyNimNode(x) + doAssert s.len == 0 + doAssert s.kind == nnkStmtList + + foo: + let x = 12 + echo x + +proc copyNimTree*(n: NimNode): NimNode {.magic: "NCopyNimTree", noSideEffect.} = + ## Creates a new AST node by recursively copying the node `n`. Note that + ## unlike `copyNimNode`, this copies `n`, the children of `n`, etc. + runnableExamples: + macro foo(x: typed) = + var s = copyNimTree(x) + doAssert s.len == 2 + doAssert s.kind == nnkStmtList -proc error*(msg: string, n: NimNode = nil) {.magic: "NError", benign.} - ## Writes an error message at compile time. The optional ``n: NimNode`` + foo: + let x = 12 + echo x + +when defined(nimHasNoReturnError): + {.pragma: errorNoReturn, noreturn.} +else: + {.pragma: errorNoReturn.} + +proc error*(msg: string, n: NimNode = nil) {.magic: "NError", benign, errorNoReturn.} + ## Writes an error message at compile time. The optional `n: NimNode` ## parameter is used as the source for file and line number information in ## the compilation error message. @@ -425,29 +444,29 @@ proc warning*(msg: string, n: NimNode = nil) {.magic: "NWarning", benign.} proc hint*(msg: string, n: NimNode = nil) {.magic: "NHint", benign.} ## Writes a hint message at compile time. -proc newStrLitNode*(s: string): NimNode {.compileTime, noSideEffect.} = +proc newStrLitNode*(s: string): NimNode {.noSideEffect.} = ## Creates a string literal node from `s`. result = newNimNode(nnkStrLit) result.strVal = s -proc newCommentStmtNode*(s: string): NimNode {.compileTime, noSideEffect.} = +proc newCommentStmtNode*(s: string): NimNode {.noSideEffect.} = ## Creates a comment statement node. result = newNimNode(nnkCommentStmt) result.strVal = s -proc newIntLitNode*(i: BiggestInt): NimNode {.compileTime.} = +proc newIntLitNode*(i: BiggestInt): NimNode = ## Creates an int literal node from `i`. result = newNimNode(nnkIntLit) result.intVal = i -proc newFloatLitNode*(f: BiggestFloat): NimNode {.compileTime.} = +proc newFloatLitNode*(f: BiggestFloat): NimNode = ## Creates a float literal node from `f`. result = newNimNode(nnkFloatLit) result.floatVal = f {.push warnings: off.} -proc newIdentNode*(i: NimIdent): NimNode {.compileTime, deprecated.} = +proc newIdentNode*(i: NimIdent): NimNode {.deprecated: "use ident(string)".} = ## Creates an identifier node from `i`. result = newNimNode(nnkIdent) result.ident = i @@ -456,15 +475,17 @@ proc newIdentNode*(i: NimIdent): NimNode {.compileTime, deprecated.} = proc newIdentNode*(i: string): NimNode {.magic: "StrToIdent", noSideEffect.} ## Creates an identifier node from `i`. It is simply an alias for - ## ``ident(string)``. Use that, it's shorter. + ## `ident(string)`. Use that, it's shorter. proc ident*(name: string): NimNode {.magic: "StrToIdent", noSideEffect.} ## Create a new ident node from a string. type - BindSymRule* = enum ## specifies how ``bindSym`` behaves + BindSymRule* = enum ## Specifies how `bindSym` behaves. The difference + ## between open and closed symbols can be found in + ## `<manual.html#symbol-lookup-in-generics-open-and-closed-symbols>`_ brClosed, ## only the symbols in current scope are bound - brOpen, ## open wrt overloaded symbols, but may be a single + brOpen, ## open for overloaded symbols, but may be a single ## symbol if not ambiguous (the rules match that of ## binding in generics) brForceOpen ## same as brOpen, but it will always be open even @@ -475,20 +496,15 @@ proc bindSym*(ident: string | NimNode, rule: BindSymRule = brClosed): NimNode {. magic: "NBindSym", noSideEffect.} ## Creates a node that binds `ident` to a symbol node. The bound symbol ## may be an overloaded symbol. - ## if `ident` is a NimNode, it must have ``nnkIdent`` kind. - ## If ``rule == brClosed`` either an ``nnkClosedSymChoice`` tree is - ## returned or ``nnkSym`` if the symbol is not ambiguous. - ## If ``rule == brOpen`` either an ``nnkOpenSymChoice`` tree is - ## returned or ``nnkSym`` if the symbol is not ambiguous. - ## If ``rule == brForceOpen`` always an ``nnkOpenSymChoice`` tree is + ## if `ident` is a NimNode, it must have `nnkIdent` kind. + ## If `rule == brClosed` either an `nnkClosedSymChoice` tree is + ## returned or `nnkSym` if the symbol is not ambiguous. + ## If `rule == brOpen` either an `nnkOpenSymChoice` tree is + ## returned or `nnkSym` if the symbol is not ambiguous. + ## If `rule == brForceOpen` always an `nnkOpenSymChoice` tree is ## returned even if the symbol is not ambiguous. ## - ## Experimental feature: - ## use {.experimental: "dynamicBindSym".} to activate it. - ## If called from template / regular code, `ident` and `rule` must be - ## constant expression / literal value. - ## If called from macros / compile time procs / static blocks, - ## `ident` and `rule` can be VM computed value. + ## See the `manual <manual.html#macros-bindsym>`_ for more details. proc genSym*(kind: NimSymKind = nskLet; ident = ""): NimNode {. magic: "NGenSym", noSideEffect.} @@ -496,10 +512,11 @@ proc genSym*(kind: NimSymKind = nskLet; ident = ""): NimNode {. ## needs to occur in a declaration context. proc callsite*(): NimNode {.magic: "NCallSite", benign, deprecated: - "Deprecated since v0.18.1; use varargs[untyped] in the macro prototype instead".} + "Deprecated since v0.18.1; use `varargs[untyped]` in the macro prototype instead".} ## Returns the AST of the invocation expression that invoked this macro. + # see https://github.com/nim-lang/RFCs/issues/387 as candidate replacement. -proc toStrLit*(n: NimNode): NimNode {.compileTime.} = +proc toStrLit*(n: NimNode): NimNode = ## Converts the AST `n` to the concrete Nim code and wraps that ## in a string literal node. return newStrLitNode(repr(n)) @@ -523,123 +540,194 @@ proc getColumn(arg: NimNode): int {.magic: "NLineInfo", noSideEffect.} proc getFile(arg: NimNode): string {.magic: "NLineInfo", noSideEffect.} proc copyLineInfo*(arg: NimNode, info: NimNode) {.magic: "NLineInfo", noSideEffect.} - ## Copy lineinfo from ``info``. - -proc lineInfoObj*(n: NimNode): LineInfo {.compileTime.} = - ## Returns ``LineInfo`` of ``n``, using absolute path for ``filename``. + ## Copy lineinfo from `info`. + +proc setLine(arg: NimNode, line: uint16) {.magic: "NLineInfo", noSideEffect.} +proc setColumn(arg: NimNode, column: int16) {.magic: "NLineInfo", noSideEffect.} +proc setFile(arg: NimNode, file: string) {.magic: "NLineInfo", noSideEffect.} + +proc setLineInfo*(arg: NimNode, file: string, line: int, column: int) = + ## Sets the line info on the NimNode. The file needs to exists, but can be a + ## relative path. If you want to attach line info to a block using `quote` + ## you'll need to add the line information after the quote block. + arg.setFile(file) + arg.setLine(line.uint16) + arg.setColumn(column.int16) + +proc setLineInfo*(arg: NimNode, lineInfo: LineInfo) = + ## See `setLineInfo proc<#setLineInfo,NimNode,string,int,int>`_ + setLineInfo(arg, lineInfo.filename, lineInfo.line, lineInfo.column) + +proc lineInfoObj*(n: NimNode): LineInfo = + ## Returns `LineInfo` of `n`, using absolute path for `filename`. result = LineInfo(filename: n.getFile, line: n.getLine, column: n.getColumn) -proc lineInfo*(arg: NimNode): string {.compileTime.} = +proc lineInfo*(arg: NimNode): string = ## Return line info in the form `filepath(line, column)`. $arg.lineInfoObj -proc internalParseExpr(s: string): NimNode {. +proc internalParseExpr(s, filename: string): NimNode {. magic: "ParseExprToAst", noSideEffect.} -proc internalParseStmt(s: string): NimNode {. +proc internalParseStmt(s, filename: string): NimNode {. magic: "ParseStmtToAst", noSideEffect.} proc internalErrorFlag*(): string {.magic: "NError", noSideEffect.} ## Some builtins set an error flag. This is then turned into a proper ## exception. **Note**: Ordinary application code should not call this. -proc parseExpr*(s: string): NimNode {.noSideEffect, compileTime.} = +proc parseExpr*(s: string; filename: string = ""): NimNode {.noSideEffect.} = ## Compiles the passed string to its AST representation. - ## Expects a single expression. Raises ``ValueError`` for parsing errors. - result = internalParseExpr(s) + ## Expects a single expression. Raises `ValueError` for parsing errors. + ## A filename can be given for more informative errors. + result = internalParseExpr(s, filename) let x = internalErrorFlag() if x.len > 0: raise newException(ValueError, x) -proc parseStmt*(s: string): NimNode {.noSideEffect, compileTime.} = +proc parseStmt*(s: string; filename: string = ""): NimNode {.noSideEffect.} = ## Compiles the passed string to its AST representation. - ## Expects one or more statements. Raises ``ValueError`` for parsing errors. - result = internalParseStmt(s) + ## Expects one or more statements. Raises `ValueError` for parsing errors. + ## A filename can be given for more informative errors. + result = internalParseStmt(s, filename) let x = internalErrorFlag() if x.len > 0: raise newException(ValueError, x) proc getAst*(macroOrTemplate: untyped): NimNode {.magic: "ExpandToAst", noSideEffect.} ## Obtains the AST nodes returned from a macro or template invocation. + ## See also `genasts.genAst`. ## Example: - ## - ## .. code-block:: nim - ## + ## ```nim ## macro FooMacro() = ## var ast = getAst(BarTemplate()) + ## ``` -proc quote*(bl: typed, op = "``"): NimNode {.magic: "QuoteAst", noSideEffect.} +proc quote*(bl: typed, op = "``"): NimNode {.magic: "QuoteAst", noSideEffect.} = ## Quasi-quoting operator. ## Accepts an expression or a block and returns the AST that represents it. ## Within the quoted AST, you are able to interpolate NimNode expressions ## from the surrounding scope. If no operator is given, quoting is done using ## backticks. Otherwise, the given operator must be used as a prefix operator - ## for any interpolated expression. - ## - ## Example: - ## - ## .. code-block:: nim + ## for any interpolated expression. The original meaning of the interpolation + ## operator may be obtained by escaping it (by prefixing it with itself) when used + ## as a unary operator: + ## e.g. `@` is escaped as `@@`, `&%` is escaped as `&%&%` and so on; see examples. ## - ## macro check(ex: untyped) = - ## # this is a simplified version of the check macro from the - ## # unittest module. + ## A custom operator interpolation needs accent quoted (``) whenever it resolves + ## to a symbol. ## - ## # If there is a failed check, we want to make it easy for - ## # the user to jump to the faulty line in the code, so we - ## # get the line info here: - ## var info = ex.lineinfo - ## - ## # We will also display the code string of the failed check: - ## var expString = ex.toStrLit - ## - ## # Finally we compose the code to implement the check: - ## result = quote do: - ## if not `ex`: - ## echo `info` & ": Check failed: " & `expString` + ## See also `genasts <genasts.html>`_ which avoids some issues with `quote`. + runnableExamples: + macro check(ex: untyped) = + # this is a simplified version of the check macro from the + # unittest module. -proc expectKind*(n: NimNode, k: NimNodeKind) {.compileTime.} = + # If there is a failed check, we want to make it easy for + # the user to jump to the faulty line in the code, so we + # get the line info here: + var info = ex.lineinfo + + # We will also display the code string of the failed check: + var expString = ex.toStrLit + + # Finally we compose the code to implement the check: + result = quote do: + if not `ex`: + echo `info` & ": Check failed: " & `expString` + check 1 + 1 == 2 + + runnableExamples: + # example showing how to define a symbol that requires backtick without + # quoting it. + var destroyCalled = false + macro bar() = + let s = newTree(nnkAccQuoted, ident"=destroy") + # let s = ident"`=destroy`" # this would not work + result = quote do: + type Foo = object + # proc `=destroy`(a: var Foo) = destroyCalled = true # this would not work + proc `s`(a: var Foo) = destroyCalled = true + block: + let a = Foo() + bar() + doAssert destroyCalled + + runnableExamples: + # custom `op` + var destroyCalled = false + macro bar(ident) = + var x = 1.5 + result = quote("@") do: + type Foo = object + let `@ident` = 0 # custom op interpolated symbols need quoted (``) + proc `=destroy`(a: var Foo) = + doAssert @x == 1.5 + doAssert compiles(@x == 1.5) + let b1 = @[1,2] + let b2 = @@[1,2] + doAssert $b1 == "[1, 2]" + doAssert $b2 == "@[1, 2]" + destroyCalled = true + block: + let a = Foo() + bar(someident) + doAssert destroyCalled + + proc `&%`(x: int): int = 1 + proc `&%`(x, y: int): int = 2 + + macro bar2() = + var x = 3 + result = quote("&%") do: + var y = &%x # quoting operator + doAssert &%&%y == 1 # unary operator => need to escape + doAssert y &% y == 2 # binary operator => no need to escape + doAssert y == 3 + bar2() + +proc expectKind*(n: NimNode, k: NimNodeKind) = ## Checks that `n` is of kind `k`. If this is not the case, ## compilation aborts with an error message. This is useful for writing ## macros that check the AST that is passed to them. if n.kind != k: error("Expected a node of kind " & $k & ", got " & $n.kind, n) -proc expectMinLen*(n: NimNode, min: int) {.compileTime.} = +proc expectMinLen*(n: NimNode, min: int) = ## Checks that `n` has at least `min` children. If this is not the case, ## compilation aborts with an error message. This is useful for writing ## macros that check its number of arguments. if n.len < min: error("Expected a node with at least " & $min & " children, got " & $n.len, n) -proc expectLen*(n: NimNode, len: int) {.compileTime.} = +proc expectLen*(n: NimNode, len: int) = ## Checks that `n` has exactly `len` children. If this is not the case, ## compilation aborts with an error message. This is useful for writing ## macros that check its number of arguments. if n.len != len: error("Expected a node with " & $len & " children, got " & $n.len, n) -proc expectLen*(n: NimNode, min, max: int) {.compileTime.} = - ## Checks that `n` has a number of children in the range ``min..max``. +proc expectLen*(n: NimNode, min, max: int) = + ## Checks that `n` has a number of children in the range `min..max`. ## If this is not the case, compilation aborts with an error message. ## This is useful for writing macros that check its number of arguments. if n.len < min or n.len > max: error("Expected a node with " & $min & ".." & $max & " children, got " & $n.len, n) proc newTree*(kind: NimNodeKind, - children: varargs[NimNode]): NimNode {.compileTime.} = + children: varargs[NimNode]): NimNode = ## Produces a new node with children. result = newNimNode(kind) result.add(children) -proc newCall*(theProc: NimNode, - args: varargs[NimNode]): NimNode {.compileTime.} = +proc newCall*(theProc: NimNode, args: varargs[NimNode]): NimNode = ## Produces a new call node. `theProc` is the proc that is called with - ## the arguments ``args[0..]``. + ## the arguments `args[0..]`. result = newNimNode(nnkCall) result.add(theProc) result.add(args) {.push warnings: off.} -proc newCall*(theProc: NimIdent, args: varargs[NimNode]): NimNode {.compileTime, deprecated: +proc newCall*(theProc: NimIdent, args: varargs[NimNode]): NimNode {.deprecated: "Deprecated since v0.18.1; use 'newCall(string, ...)' or 'newCall(NimNode, ...)' instead".} = ## Produces a new call node. `theProc` is the proc that is called with - ## the arguments ``args[0..]``. + ## the arguments `args[0..]`. result = newNimNode(nnkCall) result.add(newIdentNode(theProc)) result.add(args) @@ -647,128 +735,128 @@ proc newCall*(theProc: NimIdent, args: varargs[NimNode]): NimNode {.compileTime, {.pop.} proc newCall*(theProc: string, - args: varargs[NimNode]): NimNode {.compileTime.} = + args: varargs[NimNode]): NimNode = ## Produces a new call node. `theProc` is the proc that is called with - ## the arguments ``args[0..]``. + ## the arguments `args[0..]`. result = newNimNode(nnkCall) result.add(newIdentNode(theProc)) result.add(args) -proc newLit*(c: char): NimNode {.compileTime.} = +proc newLit*(c: char): NimNode = ## Produces a new character literal node. result = newNimNode(nnkCharLit) result.intVal = ord(c) -proc newLit*(i: int): NimNode {.compileTime.} = +proc newLit*(i: int): NimNode = ## Produces a new integer literal node. result = newNimNode(nnkIntLit) result.intVal = i -proc newLit*(i: int8): NimNode {.compileTime.} = +proc newLit*(i: int8): NimNode = ## Produces a new integer literal node. result = newNimNode(nnkInt8Lit) result.intVal = i -proc newLit*(i: int16): NimNode {.compileTime.} = +proc newLit*(i: int16): NimNode = ## Produces a new integer literal node. result = newNimNode(nnkInt16Lit) result.intVal = i -proc newLit*(i: int32): NimNode {.compileTime.} = +proc newLit*(i: int32): NimNode = ## Produces a new integer literal node. result = newNimNode(nnkInt32Lit) result.intVal = i -proc newLit*(i: int64): NimNode {.compileTime.} = +proc newLit*(i: int64): NimNode = ## Produces a new integer literal node. result = newNimNode(nnkInt64Lit) result.intVal = i -proc newLit*(i: uint): NimNode {.compileTime.} = +proc newLit*(i: uint): NimNode = ## Produces a new unsigned integer literal node. result = newNimNode(nnkUIntLit) result.intVal = BiggestInt(i) -proc newLit*(i: uint8): NimNode {.compileTime.} = +proc newLit*(i: uint8): NimNode = ## Produces a new unsigned integer literal node. result = newNimNode(nnkUInt8Lit) result.intVal = BiggestInt(i) -proc newLit*(i: uint16): NimNode {.compileTime.} = +proc newLit*(i: uint16): NimNode = ## Produces a new unsigned integer literal node. result = newNimNode(nnkUInt16Lit) result.intVal = BiggestInt(i) -proc newLit*(i: uint32): NimNode {.compileTime.} = +proc newLit*(i: uint32): NimNode = ## Produces a new unsigned integer literal node. result = newNimNode(nnkUInt32Lit) result.intVal = BiggestInt(i) -proc newLit*(i: uint64): NimNode {.compileTime.} = +proc newLit*(i: uint64): NimNode = ## Produces a new unsigned integer literal node. result = newNimNode(nnkUInt64Lit) result.intVal = BiggestInt(i) -proc newLit*(b: bool): NimNode {.compileTime.} = +proc newLit*(b: bool): NimNode = ## Produces a new boolean literal node. result = if b: bindSym"true" else: bindSym"false" -proc newLit*(s: string): NimNode {.compileTime.} = +proc newLit*(s: string): NimNode = ## Produces a new string literal node. result = newNimNode(nnkStrLit) result.strVal = s when false: # the float type is not really a distinct type as described in https://github.com/nim-lang/Nim/issues/5875 - proc newLit*(f: float): NimNode {.compileTime.} = + proc newLit*(f: float): NimNode = ## Produces a new float literal node. result = newNimNode(nnkFloatLit) result.floatVal = f -proc newLit*(f: float32): NimNode {.compileTime.} = +proc newLit*(f: float32): NimNode = ## Produces a new float literal node. result = newNimNode(nnkFloat32Lit) result.floatVal = f -proc newLit*(f: float64): NimNode {.compileTime.} = +proc newLit*(f: float64): NimNode = ## Produces a new float literal node. result = newNimNode(nnkFloat64Lit) result.floatVal = f when declared(float128): - proc newLit*(f: float128): NimNode {.compileTime.} = + proc newLit*(f: float128): NimNode = ## Produces a new float literal node. result = newNimNode(nnkFloat128Lit) result.floatVal = f -proc newLit*(arg: enum): NimNode {.compileTime.} = +proc newLit*(arg: enum): NimNode = result = newCall( - arg.type.getTypeInst[1], + arg.typeof.getTypeInst[1], newLit(int(arg)) ) -proc newLit*[N,T](arg: array[N,T]): NimNode {.compileTime.} -proc newLit*[T](arg: seq[T]): NimNode {.compileTime.} -proc newLit*[T](s: set[T]): NimNode {.compileTime.} -proc newLit*[T: tuple](arg: T): NimNode {.compileTime.} +proc newLit*[N,T](arg: array[N,T]): NimNode +proc newLit*[T](arg: seq[T]): NimNode +proc newLit*[T](s: set[T]): NimNode +proc newLit*[T: tuple](arg: T): NimNode -proc newLit*(arg: object): NimNode {.compileTime.} = - result = nnkObjConstr.newTree(arg.type.getTypeInst[1]) +proc newLit*(arg: object): NimNode = + result = nnkObjConstr.newTree(arg.typeof.getTypeInst[1]) for a, b in arg.fieldPairs: result.add nnkExprColonExpr.newTree( newIdentNode(a), newLit(b) ) -proc newLit*(arg: ref object): NimNode {.compileTime.} = +proc newLit*(arg: ref object): NimNode = ## produces a new ref type literal node. - result = nnkObjConstr.newTree(arg.type.getTypeInst[1]) + result = nnkObjConstr.newTree(arg.typeof.getTypeInst[1]) for a, b in fieldPairs(arg[]): result.add nnkExprColonExpr.newTree(newIdentNode(a), newLit(b)) -proc newLit*[N,T](arg: array[N,T]): NimNode {.compileTime.} = +proc newLit*[N,T](arg: array[N,T]): NimNode = result = nnkBracket.newTree for x in arg: result.add newLit(x) -proc newLit*[T](arg: seq[T]): NimNode {.compileTime.} = +proc newLit*[T](arg: seq[T]): NimNode = let bracket = nnkBracket.newTree for x in arg: bracket.add newLit(x) @@ -781,7 +869,7 @@ proc newLit*[T](arg: seq[T]): NimNode {.compileTime.} = var typ = getTypeInst(typeof(arg))[1] result = newCall(typ,result) -proc newLit*[T](s: set[T]): NimNode {.compileTime.} = +proc newLit*[T](s: set[T]): NimNode = result = nnkCurly.newTree for x in s: result.add newLit(x) @@ -793,7 +881,7 @@ proc newLit*[T](s: set[T]): NimNode {.compileTime.} = proc isNamedTuple(T: typedesc): bool {.magic: "TypeTrait".} ## See `typetraits.isNamedTuple` -proc newLit*[T: tuple](arg: T): NimNode {.compileTime.} = +proc newLit*[T: tuple](arg: T): NimNode = ## use -d:nimHasWorkaround14720 to restore behavior prior to PR, forcing ## a named tuple even when `arg` is unnamed. result = nnkTupleConstr.newTree @@ -804,9 +892,9 @@ proc newLit*[T: tuple](arg: T): NimNode {.compileTime.} = for b in arg.fields: result.add newLit(b) -proc nestList*(op: NimNode; pack: NimNode): NimNode {.compileTime.} = +proc nestList*(op: NimNode; pack: NimNode): NimNode = ## Nests the list `pack` into a tree of call expressions: - ## ``[a, b, c]`` is transformed into ``op(a, op(c, d))``. + ## `[a, b, c]` is transformed into `op(a, op(c, d))`. ## This is also known as fold expression. if pack.len < 1: error("`nestList` expects a node with at least 1 child") @@ -814,14 +902,37 @@ proc nestList*(op: NimNode; pack: NimNode): NimNode {.compileTime.} = for i in countdown(pack.len - 2, 0): result = newCall(op, pack[i], result) -proc nestList*(op: NimNode; pack: NimNode; init: NimNode): NimNode {.compileTime.} = +proc nestList*(op: NimNode; pack: NimNode; init: NimNode): NimNode = ## Nests the list `pack` into a tree of call expressions: - ## ``[a, b, c]`` is transformed into ``op(a, op(c, d))``. + ## `[a, b, c]` is transformed into `op(a, op(c, d))`. ## This is also known as fold expression. result = init for i in countdown(pack.len - 1, 0): result = newCall(op, pack[i], result) +proc eqIdent*(a: string; b: string): bool {.magic: "EqIdent", noSideEffect.} + ## Style insensitive comparison. + +proc eqIdent*(a: NimNode; b: string): bool {.magic: "EqIdent", noSideEffect.} + ## Style insensitive comparison. `a` can be an identifier or a + ## symbol. `a` may be wrapped in an export marker + ## (`nnkPostfix`) or quoted with backticks (`nnkAccQuoted`), + ## these nodes will be unwrapped. + +proc eqIdent*(a: string; b: NimNode): bool {.magic: "EqIdent", noSideEffect.} + ## Style insensitive comparison. `b` can be an identifier or a + ## symbol. `b` may be wrapped in an export marker + ## (`nnkPostfix`) or quoted with backticks (`nnkAccQuoted`), + ## these nodes will be unwrapped. + +proc eqIdent*(a: NimNode; b: NimNode): bool {.magic: "EqIdent", noSideEffect.} + ## Style insensitive comparison. `a` and `b` can be an + ## identifier or a symbol. Both may be wrapped in an export marker + ## (`nnkPostfix`) or quoted with backticks (`nnkAccQuoted`), + ## these nodes will be unwrapped. + +const collapseSymChoice = not defined(nimLegacyMacrosCollapseSymChoice) + proc treeTraverse(n: NimNode; res: var string; level = 0; isLisp = false, indented = false) {.benign.} = if level > 0: if indented: @@ -843,37 +954,52 @@ proc treeTraverse(n: NimNode; res: var string; level = 0; isLisp = false, indent discard # same as nil node in this representation of nnkCharLit .. nnkInt64Lit: res.add(" " & $n.intVal) + of nnkUIntLit .. nnkUInt64Lit: + res.add(" " & $cast[uint64](n.intVal)) of nnkFloatLit .. nnkFloat64Lit: res.add(" " & $n.floatVal) of nnkStrLit .. nnkTripleStrLit, nnkCommentStmt, nnkIdent, nnkSym: res.add(" " & $n.strVal.newLit.repr) of nnkNone: assert false + elif n.kind in {nnkOpenSymChoice, nnkClosedSymChoice} and collapseSymChoice: + res.add(" " & $n.len) + if n.len > 0: + var allSameSymName = true + for i in 0..<n.len: + if n[i].kind != nnkSym or not eqIdent(n[i], n[0]): + allSameSymName = false + break + if allSameSymName: + res.add(" " & $n[0].strVal.newLit.repr) + else: + for j in 0 ..< n.len: + n[j].treeTraverse(res, level+1, isLisp, indented) else: - for j in 0 .. n.len-1: + for j in 0 ..< n.len: n[j].treeTraverse(res, level+1, isLisp, indented) if isLisp: res.add(")") -proc treeRepr*(n: NimNode): string {.compileTime, benign.} = +proc treeRepr*(n: NimNode): string {.benign.} = ## Convert the AST `n` to a human-readable tree-like string. ## - ## See also `repr`, `lispRepr`, and `astGenRepr`. + ## See also `repr`, `lispRepr`_, and `astGenRepr`_. result = "" n.treeTraverse(result, isLisp = false, indented = true) -proc lispRepr*(n: NimNode; indented = false): string {.compileTime, benign.} = - ## Convert the AST ``n`` to a human-readable lisp-like string. +proc lispRepr*(n: NimNode; indented = false): string {.benign.} = + ## Convert the AST `n` to a human-readable lisp-like string. ## - ## See also ``repr``, ``treeRepr``, and ``astGenRepr``. + ## See also `repr`, `treeRepr`_, and `astGenRepr`_. result = "" n.treeTraverse(result, isLisp = true, indented = indented) -proc astGenRepr*(n: NimNode): string {.compileTime, benign.} = - ## Convert the AST ``n`` to the code required to generate that AST. +proc astGenRepr*(n: NimNode): string {.benign.} = + ## Convert the AST `n` to the code required to generate that AST. ## - ## See also ``repr``, ``treeRepr``, and ``lispRepr``. + ## See also `repr`_, `treeRepr`_, and `lispRepr`_. const NodeKinds = {nnkEmpty, nnkIdent, nnkSym, nnkNone, nnkCommentStmt} @@ -898,6 +1024,10 @@ proc astGenRepr*(n: NimNode): string {.compileTime, benign.} = of nnkStrLit..nnkTripleStrLit, nnkCommentStmt, nnkIdent, nnkSym: res.add(n.strVal.newLit.repr) of nnkNone: assert false + elif n.kind in {nnkOpenSymChoice, nnkClosedSymChoice} and collapseSymChoice: + res.add(", # unrepresentable symbols: " & $n.len) + if n.len > 0: + res.add(" " & n[0].strVal.newLit.repr) else: res.add(".newTree(") for j in 0..<n.len: @@ -918,162 +1048,166 @@ proc astGenRepr*(n: NimNode): string {.compileTime, benign.} = macro dumpTree*(s: untyped): untyped = echo s.treeRepr ## Accepts a block of nim code and prints the parsed abstract syntax - ## tree using the ``treeRepr`` proc. Printing is done *at compile time*. + ## tree using the `treeRepr` proc. Printing is done *at compile time*. ## ## You can use this as a tool to explore the Nim's abstract syntax ## tree and to discover what kind of nodes must be created to represent ## a certain expression/statement. ## ## For example: - ## - ## .. code-block:: nim - ## dumpTree: - ## echo "Hello, World!" + ## ```nim + ## dumpTree: + ## echo "Hello, World!" + ## ``` ## ## Outputs: + ## ``` + ## StmtList + ## Command + ## Ident "echo" + ## StrLit "Hello, World!" + ## ``` ## - ## .. code-block:: - ## StmtList - ## Command - ## Ident "echo" - ## StrLit "Hello, World!" - ## - ## Also see ``dumpAstGen`` and ``dumpLisp``. + ## Also see `dumpAstGen` and `dumpLisp`. macro dumpLisp*(s: untyped): untyped = echo s.lispRepr(indented = true) ## Accepts a block of nim code and prints the parsed abstract syntax - ## tree using the ``lispRepr`` proc. Printing is done *at compile time*. + ## tree using the `lispRepr` proc. Printing is done *at compile time*. ## ## You can use this as a tool to explore the Nim's abstract syntax ## tree and to discover what kind of nodes must be created to represent ## a certain expression/statement. ## ## For example: - ## - ## .. code-block:: nim - ## dumpLisp: - ## echo "Hello, World!" + ## ```nim + ## dumpLisp: + ## echo "Hello, World!" + ## ``` ## ## Outputs: + ## ``` + ## (StmtList + ## (Command + ## (Ident "echo") + ## (StrLit "Hello, World!"))) + ## ``` ## - ## .. code-block:: - ## (StmtList - ## (Command - ## (Ident "echo") - ## (StrLit "Hello, World!"))) - ## - ## Also see ``dumpAstGen`` and ``dumpTree``. + ## Also see `dumpAstGen` and `dumpTree`. macro dumpAstGen*(s: untyped): untyped = echo s.astGenRepr ## Accepts a block of nim code and prints the parsed abstract syntax - ## tree using the ``astGenRepr`` proc. Printing is done *at compile time*. + ## tree using the `astGenRepr` proc. Printing is done *at compile time*. ## ## You can use this as a tool to write macros quicker by writing example ## outputs and then copying the snippets into the macro for modification. ## ## For example: - ## - ## .. code-block:: nim - ## dumpAstGen: - ## echo "Hello, World!" + ## ```nim + ## dumpAstGen: + ## echo "Hello, World!" + ## ``` ## ## Outputs: - ## - ## .. code-block:: nim - ## nnkStmtList.newTree( - ## nnkCommand.newTree( - ## newIdentNode("echo"), - ## newLit("Hello, World!") - ## ) - ## ) - ## - ## Also see ``dumpTree`` and ``dumpLisp``. - -proc newEmptyNode*(): NimNode {.compileTime, noSideEffect.} = + ## ``` + ## nnkStmtList.newTree( + ## nnkCommand.newTree( + ## newIdentNode("echo"), + ## newLit("Hello, World!") + ## ) + ## ) + ## ``` + ## + ## Also see `dumpTree` and `dumpLisp`. + +proc newEmptyNode*(): NimNode {.noSideEffect.} = ## Create a new empty node. result = newNimNode(nnkEmpty) -proc newStmtList*(stmts: varargs[NimNode]): NimNode {.compileTime.}= +proc newStmtList*(stmts: varargs[NimNode]): NimNode = ## Create a new statement list. result = newNimNode(nnkStmtList).add(stmts) -proc newPar*(exprs: varargs[NimNode]): NimNode {.compileTime.}= +proc newPar*(exprs: NimNode): NimNode = ## Create a new parentheses-enclosed expression. newNimNode(nnkPar).add(exprs) -proc newBlockStmt*(label, body: NimNode): NimNode {.compileTime.} = +proc newPar*(exprs: varargs[NimNode]): NimNode {.deprecated: + "don't use newPar/nnkPar to construct tuple expressions; use nnkTupleConstr instead".} = + ## Create a new parentheses-enclosed expression. + newNimNode(nnkPar).add(exprs) + +proc newBlockStmt*(label, body: NimNode): NimNode = ## Create a new block statement with label. return newNimNode(nnkBlockStmt).add(label, body) -proc newBlockStmt*(body: NimNode): NimNode {.compileTime.} = +proc newBlockStmt*(body: NimNode): NimNode = ## Create a new block: stmt. return newNimNode(nnkBlockStmt).add(newEmptyNode(), body) -proc newVarStmt*(name, value: NimNode): NimNode {.compileTime.} = +proc newVarStmt*(name, value: NimNode): NimNode = ## Create a new var stmt. return newNimNode(nnkVarSection).add( newNimNode(nnkIdentDefs).add(name, newNimNode(nnkEmpty), value)) -proc newLetStmt*(name, value: NimNode): NimNode {.compileTime.} = +proc newLetStmt*(name, value: NimNode): NimNode = ## Create a new let stmt. return newNimNode(nnkLetSection).add( newNimNode(nnkIdentDefs).add(name, newNimNode(nnkEmpty), value)) -proc newConstStmt*(name, value: NimNode): NimNode {.compileTime.} = +proc newConstStmt*(name, value: NimNode): NimNode = ## Create a new const stmt. newNimNode(nnkConstSection).add( newNimNode(nnkConstDef).add(name, newNimNode(nnkEmpty), value)) -proc newAssignment*(lhs, rhs: NimNode): NimNode {.compileTime.} = +proc newAssignment*(lhs, rhs: NimNode): NimNode = return newNimNode(nnkAsgn).add(lhs, rhs) -proc newDotExpr*(a, b: NimNode): NimNode {.compileTime.} = +proc newDotExpr*(a, b: NimNode): NimNode = ## Create new dot expression. ## a.dot(b) -> `a.b` return newNimNode(nnkDotExpr).add(a, b) -proc newColonExpr*(a, b: NimNode): NimNode {.compileTime.} = +proc newColonExpr*(a, b: NimNode): NimNode = ## Create new colon expression. ## newColonExpr(a, b) -> `a: b` newNimNode(nnkExprColonExpr).add(a, b) proc newIdentDefs*(name, kind: NimNode; - default = newEmptyNode()): NimNode {.compileTime.} = - ## Creates a new ``nnkIdentDefs`` node of a specific kind and value. + default = newEmptyNode()): NimNode = + ## Creates a new `nnkIdentDefs` node of a specific kind and value. ## - ## ``nnkIdentDefs`` need to have at least three children, but they can have + ## `nnkIdentDefs` need to have at least three children, but they can have ## more: first comes a list of identifiers followed by a type and value ## nodes. This helper proc creates a three node subtree, the first subnode - ## being a single identifier name. Both the ``kind`` node and ``default`` - ## (value) nodes may be empty depending on where the ``nnkIdentDefs`` - ## appears: tuple or object definitions will have an empty ``default`` node, - ## ``let`` or ``var`` blocks may have an empty ``kind`` node if the + ## being a single identifier name. Both the `kind` node and `default` + ## (value) nodes may be empty depending on where the `nnkIdentDefs` + ## appears: tuple or object definitions will have an empty `default` node, + ## `let` or `var` blocks may have an empty `kind` node if the ## identifier is being assigned a value. Example: ## - ## .. code-block:: nim - ## + ## ```nim ## var varSection = newNimNode(nnkVarSection).add( ## newIdentDefs(ident("a"), ident("string")), ## newIdentDefs(ident("b"), newEmptyNode(), newLit(3))) ## # --> var ## # a: string ## # b = 3 + ## ``` ## ## If you need to create multiple identifiers you need to use the lower level - ## ``newNimNode``: - ## - ## .. code-block:: nim - ## + ## `newNimNode`: + ## ```nim ## result = newNimNode(nnkIdentDefs).add( ## ident("a"), ident("b"), ident("c"), ident("string"), ## newStrLitNode("Hello")) + ## ``` newNimNode(nnkIdentDefs).add(name, kind, default) -proc newNilLit*(): NimNode {.compileTime.} = +proc newNilLit*(): NimNode = ## New nil literal shortcut. result = newNimNode(nnkNilLit) -proc last*(node: NimNode): NimNode {.compileTime.} = node[node.len-1] +proc last*(node: NimNode): NimNode = node[node.len-1] ## Return the last item in nodes children. Same as `node[^1]`. @@ -1081,10 +1215,10 @@ const RoutineNodes* = {nnkProcDef, nnkFuncDef, nnkMethodDef, nnkDo, nnkLambda, nnkIteratorDef, nnkTemplateDef, nnkConverterDef, nnkMacroDef} AtomicNodes* = {nnkNone..nnkNilLit} - CallNodes* = {nnkCall, nnkInfix, nnkPrefix, nnkPostfix, nnkCommand, - nnkCallStrLit, nnkHiddenCallConv} + # see matching set nnkCallKinds above + CallNodes* = nnkCallKinds -proc expectKind*(n: NimNode; k: set[NimNodeKind]) {.compileTime.} = +proc expectKind*(n: NimNode; k: set[NimNodeKind]) = ## Checks that `n` is of kind `k`. If this is not the case, ## compilation aborts with an error message. This is useful for writing ## macros that check the AST that is passed to them. @@ -1094,10 +1228,10 @@ proc newProc*(name = newEmptyNode(); params: openArray[NimNode] = [newEmptyNode()]; body: NimNode = newStmtList(); procType = nnkProcDef; - pragmas: NimNode = newEmptyNode()): NimNode {.compileTime.} = + pragmas: NimNode = newEmptyNode()): NimNode = ## Shortcut for creating a new proc. ## - ## The ``params`` array must start with the return type of the proc, + ## The `params` array must start with the return type of the proc, ## followed by a list of IdentDefs which specify the params. if procType notin RoutineNodes: error("Expected one of " & $RoutineNodes & ", got " & $procType) @@ -1111,17 +1245,14 @@ proc newProc*(name = newEmptyNode(); newEmptyNode(), body) -proc newIfStmt*(branches: varargs[tuple[cond, body: NimNode]]): - NimNode {.compileTime.} = - ## Constructor for ``if`` statements. - ## - ## .. code-block:: nim - ## - ## newIfStmt( - ## (Ident, StmtList), - ## ... - ## ) - ## +proc newIfStmt*(branches: varargs[tuple[cond, body: NimNode]]): NimNode = + ## Constructor for `if` statements. + ## ```nim + ## newIfStmt( + ## (Ident, StmtList), + ## ... + ## ) + ## ``` result = newNimNode(nnkIfStmt) if len(branches) < 1: error("If statement must have at least one branch") @@ -1129,20 +1260,18 @@ proc newIfStmt*(branches: varargs[tuple[cond, body: NimNode]]): result.add(newTree(nnkElifBranch, i.cond, i.body)) proc newEnum*(name: NimNode, fields: openArray[NimNode], - public, pure: bool): NimNode {.compileTime.} = + public, pure: bool): NimNode = ## Creates a new enum. `name` must be an ident. Fields are allowed to be - ## either idents or EnumFieldDef - ## - ## .. code-block:: nim - ## - ## newEnum( - ## name = ident("Colors"), - ## fields = [ident("Blue"), ident("Red")], - ## public = true, pure = false) - ## - ## # type Colors* = Blue Red + ## either idents or EnumFieldDef: + ## ```nim + ## newEnum( + ## name = ident("Colors"), + ## fields = [ident("Blue"), ident("Red")], + ## public = true, pure = false) ## + ## # type Colors* = Blue Red + ## ``` expectKind name, nnkIdent if len(fields) < 1: @@ -1172,7 +1301,7 @@ proc newEnum*(name: NimNode, fields: openArray[NimNode], return typeSect -proc copyChildrenTo*(src, dest: NimNode) {.compileTime.}= +proc copyChildrenTo*(src, dest: NimNode) = ## Copy all children from `src` to `dest`. for i in 0 ..< src.len: dest.add src[i].copyNimTree @@ -1180,7 +1309,7 @@ proc copyChildrenTo*(src, dest: NimNode) {.compileTime.}= template expectRoutine(node: NimNode) = expectKind(node, RoutineNodes) -proc name*(someProc: NimNode): NimNode {.compileTime.} = +proc name*(someProc: NimNode): NimNode = someProc.expectRoutine result = someProc[0] if result.kind == nnkPostfix: @@ -1191,21 +1320,28 @@ proc name*(someProc: NimNode): NimNode {.compileTime.} = elif result.kind == nnkAccQuoted: result = result[0] -proc `name=`*(someProc: NimNode; val: NimNode) {.compileTime.} = +proc `name=`*(someProc: NimNode; val: NimNode) = someProc.expectRoutine if someProc[0].kind == nnkPostfix: someProc[0][1] = val else: someProc[0] = val -proc params*(someProc: NimNode): NimNode {.compileTime.} = - someProc.expectRoutine - result = someProc[3] -proc `params=`* (someProc: NimNode; params: NimNode) {.compileTime.}= - someProc.expectRoutine +proc params*(someProc: NimNode): NimNode = + if someProc.kind == nnkProcTy: + someProc[0] + else: + someProc.expectRoutine + someProc[3] + +proc `params=`* (someProc: NimNode; params: NimNode) = expectKind(params, nnkFormalParams) - someProc[3] = params + if someProc.kind == nnkProcTy: + someProc[0] = params + else: + someProc.expectRoutine + someProc[3] = params -proc pragma*(someProc: NimNode): NimNode {.compileTime.} = +proc pragma*(someProc: NimNode): NimNode = ## Get the pragma of a proc type. ## These will be expanded. if someProc.kind == nnkProcTy: @@ -1213,7 +1349,7 @@ proc pragma*(someProc: NimNode): NimNode {.compileTime.} = else: someProc.expectRoutine result = someProc[4] -proc `pragma=`*(someProc: NimNode; val: NimNode) {.compileTime.}= +proc `pragma=`*(someProc: NimNode; val: NimNode) = ## Set the pragma of a proc type. expectKind(val, {nnkEmpty, nnkPragma}) if someProc.kind == nnkProcTy: @@ -1222,7 +1358,7 @@ proc `pragma=`*(someProc: NimNode; val: NimNode) {.compileTime.}= someProc.expectRoutine someProc[4] = val -proc addPragma*(someProc, pragma: NimNode) {.compileTime.} = +proc addPragma*(someProc, pragma: NimNode) = ## Adds pragma to routine definition. someProc.expectKind(RoutineNodes + {nnkProcTy}) var pragmaNode = someProc.pragma @@ -1234,7 +1370,7 @@ proc addPragma*(someProc, pragma: NimNode) {.compileTime.} = template badNodeKind(n, f) = error("Invalid node kind " & $n.kind & " for macros.`" & $f & "`", n) -proc body*(someProc: NimNode): NimNode {.compileTime.} = +proc body*(someProc: NimNode): NimNode = case someProc.kind: of RoutineNodes: return someProc[6] @@ -1245,7 +1381,7 @@ proc body*(someProc: NimNode): NimNode {.compileTime.} = else: badNodeKind someProc, "body" -proc `body=`*(someProc: NimNode, val: NimNode) {.compileTime.} = +proc `body=`*(someProc: NimNode, val: NimNode) = case someProc.kind of RoutineNodes: someProc[6] = val @@ -1256,43 +1392,53 @@ proc `body=`*(someProc: NimNode, val: NimNode) {.compileTime.} = else: badNodeKind someProc, "body=" -proc basename*(a: NimNode): NimNode {.compileTime, benign.} +proc basename*(a: NimNode): NimNode = + ## Pull an identifier from prefix/postfix expressions. + case a.kind + of nnkIdent: result = a + of nnkPostfix, nnkPrefix: result = a[1] + of nnkPragmaExpr: result = basename(a[0]) + else: + error("Do not know how to get basename of (" & treeRepr(a) & ")\n" & + repr(a), a) -proc `$`*(node: NimNode): string {.compileTime.} = +proc `$`*(node: NimNode): string = ## Get the string of an identifier node. case node.kind of nnkPostfix: result = node.basename.strVal & "*" of nnkStrLit..nnkTripleStrLit, nnkCommentStmt, nnkSym, nnkIdent: result = node.strVal - of nnkOpenSymChoice, nnkClosedSymChoice: + of nnkOpenSymChoice, nnkClosedSymChoice, nnkOpenSym: result = $node[0] of nnkAccQuoted: - result = $node[0] + result = "" + for i in 0 ..< node.len: + result.add(repr(node[i])) else: badNodeKind node, "$" iterator items*(n: NimNode): NimNode {.inline.} = - ## Iterates over the children of the NimNode ``n``. + ## Iterates over the children of the NimNode `n`. for i in 0 ..< n.len: yield n[i] iterator pairs*(n: NimNode): (int, NimNode) {.inline.} = - ## Iterates over the children of the NimNode ``n`` and its indices. + ## Iterates over the children of the NimNode `n` and its indices. for i in 0 ..< n.len: yield (i, n[i]) iterator children*(n: NimNode): NimNode {.inline.} = - ## Iterates over the children of the NimNode ``n``. + ## Iterates over the children of the NimNode `n`. for i in 0 ..< n.len: yield n[i] template findChild*(n: NimNode; cond: untyped): NimNode {.dirty.} = ## Find the first child node matching condition (or nil). - ## - ## .. code-block:: nim + ## ```nim ## var res = findChild(n, it.kind == nnkPostfix and - ## it.basename.ident == toNimIdent"foo") + ## it.basename.ident == ident"foo") + ## ``` block: var res: NimNode for it in n.children: @@ -1301,8 +1447,8 @@ template findChild*(n: NimNode; cond: untyped): NimNode {.dirty.} = break res -proc insert*(a: NimNode; pos: int; b: NimNode) {.compileTime.} = - ## Insert node ``b`` into node ``a`` at ``pos``. +proc insert*(a: NimNode; pos: int; b: NimNode) = + ## Insert node `b` into node `a` at `pos`. if len(a)-1 < pos: # add some empty nodes first for i in len(a)-1..pos-2: @@ -1316,17 +1462,7 @@ proc insert*(a: NimNode; pos: int; b: NimNode) {.compileTime.} = a[i + 1] = a[i] a[pos] = b -proc basename*(a: NimNode): NimNode = - ## Pull an identifier from prefix/postfix expressions. - case a.kind - of nnkIdent: result = a - of nnkPostfix, nnkPrefix: result = a[1] - of nnkPragmaExpr: result = basename(a[0]) - else: - error("Do not know how to get basename of (" & treeRepr(a) & ")\n" & - repr(a), a) - -proc `basename=`*(a: NimNode; val: string) {.compileTime.}= +proc `basename=`*(a: NimNode; val: string) = case a.kind of nnkIdent: a.strVal = val @@ -1337,110 +1473,49 @@ proc `basename=`*(a: NimNode; val: string) {.compileTime.}= error("Do not know how to get basename of (" & treeRepr(a) & ")\n" & repr(a), a) -proc postfix*(node: NimNode; op: string): NimNode {.compileTime.} = +proc postfix*(node: NimNode; op: string): NimNode = newNimNode(nnkPostfix).add(ident(op), node) -proc prefix*(node: NimNode; op: string): NimNode {.compileTime.} = +proc prefix*(node: NimNode; op: string): NimNode = newNimNode(nnkPrefix).add(ident(op), node) proc infix*(a: NimNode; op: string; - b: NimNode): NimNode {.compileTime.} = + b: NimNode): NimNode = newNimNode(nnkInfix).add(ident(op), a, b) -proc unpackPostfix*(node: NimNode): tuple[node: NimNode; op: string] {. - compileTime.} = +proc unpackPostfix*(node: NimNode): tuple[node: NimNode; op: string] = node.expectKind nnkPostfix result = (node[1], $node[0]) -proc unpackPrefix*(node: NimNode): tuple[node: NimNode; op: string] {. - compileTime.} = +proc unpackPrefix*(node: NimNode): tuple[node: NimNode; op: string] = node.expectKind nnkPrefix result = (node[1], $node[0]) -proc unpackInfix*(node: NimNode): tuple[left: NimNode; op: string; - right: NimNode] {.compileTime.} = +proc unpackInfix*(node: NimNode): tuple[left: NimNode; op: string; right: NimNode] = expectKind(node, nnkInfix) result = (node[1], $node[0], node[2]) -proc copy*(node: NimNode): NimNode {.compileTime.} = +proc copy*(node: NimNode): NimNode = ## An alias for `copyNimTree<#copyNimTree,NimNode>`_. return node.copyNimTree() -when defined(nimVmEqIdent): - proc eqIdent*(a: string; b: string): bool {.magic: "EqIdent", noSideEffect.} - ## Style insensitive comparison. - - proc eqIdent*(a: NimNode; b: string): bool {.magic: "EqIdent", noSideEffect.} - ## Style insensitive comparison. ``a`` can be an identifier or a - ## symbol. ``a`` may be wrapped in an export marker - ## (``nnkPostfix``) or quoted with backticks (``nnkAccQuoted``), - ## these nodes will be unwrapped. - - proc eqIdent*(a: string; b: NimNode): bool {.magic: "EqIdent", noSideEffect.} - ## Style insensitive comparison. ``b`` can be an identifier or a - ## symbol. ``b`` may be wrapped in an export marker - ## (``nnkPostfix``) or quoted with backticks (``nnkAccQuoted``), - ## these nodes will be unwrapped. - - proc eqIdent*(a: NimNode; b: NimNode): bool {.magic: "EqIdent", noSideEffect.} - ## Style insensitive comparison. ``a`` and ``b`` can be an - ## identifier or a symbol. Both may be wrapped in an export marker - ## (``nnkPostfix``) or quoted with backticks (``nnkAccQuoted``), - ## these nodes will be unwrapped. - -else: - # this procedure is optimized for native code, it should not be compiled to nimVM bytecode. - proc cmpIgnoreStyle(a, b: cstring): int {.noSideEffect.} = - proc toLower(c: char): char {.inline.} = - if c in {'A'..'Z'}: result = chr(ord(c) + (ord('a') - ord('A'))) - else: result = c - var i = 0 - var j = 0 - # first char is case sensitive - if a[0] != b[0]: return 1 - while true: - while a[i] == '_': inc(i) - while b[j] == '_': inc(j) # BUGFIX: typo - var aa = toLower(a[i]) - var bb = toLower(b[j]) - result = ord(aa) - ord(bb) - if result != 0 or aa == '\0': break - inc(i) - inc(j) - - - proc eqIdent*(a, b: string): bool = cmpIgnoreStyle(a, b) == 0 - ## Check if two idents are equal. - - proc eqIdent*(node: NimNode; s: string): bool {.compileTime.} = - ## Check if node is some identifier node (``nnkIdent``, ``nnkSym``, etc.) - ## is the same as ``s``. Note that this is the preferred way to check! Most - ## other ways like ``node.ident`` are much more error-prone, unfortunately. - case node.kind - of nnkSym, nnkIdent: - result = eqIdent(node.strVal, s) - of nnkOpenSymChoice, nnkClosedSymChoice: - result = eqIdent($node[0], s) - else: - result = false - -proc expectIdent*(n: NimNode, name: string) {.compileTime, since: (1,1).} = - ## Check that ``eqIdent(n,name)`` holds true. If this is not the +proc expectIdent*(n: NimNode, name: string) {.since: (1,1).} = + ## Check that `eqIdent(n,name)` holds true. If this is not the ## case, compilation aborts with an error message. This is useful ## for writing macros that check the AST that is passed to them. if not eqIdent(n, name): error("Expected identifier to be `" & name & "` here", n) -proc hasArgOfName*(params: NimNode; name: string): bool {.compileTime.}= - ## Search ``nnkFormalParams`` for an argument. +proc hasArgOfName*(params: NimNode; name: string): bool = + ## Search `nnkFormalParams` for an argument. expectKind(params, nnkFormalParams) for i in 1..<params.len: for j in 0..<params[i].len-2: if name.eqIdent($params[i][j]): return true -proc addIdentIfAbsent*(dest: NimNode, ident: string) {.compileTime.} = - ## Add ``ident`` to ``dest`` if it is not present. This is intended for use +proc addIdentIfAbsent*(dest: NimNode, ident: string) = + ## Add `ident` to `dest` if it is not present. This is intended for use ## with pragmas. for node in dest.children: case node.kind @@ -1451,15 +1526,14 @@ proc addIdentIfAbsent*(dest: NimNode, ident: string) {.compileTime.} = else: discard dest.add(ident(ident)) -proc boolVal*(n: NimNode): bool {.compileTime, noSideEffect.} = +proc boolVal*(n: NimNode): bool {.noSideEffect.} = if n.kind == nnkIntLit: n.intVal != 0 else: n == bindSym"true" # hacky solution for now -when defined(nimMacrosGetNodeId): - proc nodeID*(n: NimNode): int {.magic: "NodeId".} - ## Returns the id of ``n``, when the compiler has been compiled - ## with the flag ``-d:useNodeids``, otherwise returns ``-1``. This - ## proc is for the purpose to debug the compiler only. +proc nodeID*(n: NimNode): int {.magic: "NodeId".} + ## Returns the id of `n`, when the compiler has been compiled + ## with the flag `-d:useNodeids`, otherwise returns `-1`. This + ## proc is for the purpose to debug the compiler only. macro expandMacros*(body: typed): untyped = ## Expands one level of macro - useful for debugging. @@ -1468,31 +1542,53 @@ macro expandMacros*(body: typed): untyped = ## ## For instance, ## - ## .. code-block:: nim - ## import sugar, macros + ## ```nim + ## import std/[sugar, macros] ## ## let ## x = 10 ## y = 20 ## expandMacros: ## dump(x + y) + ## ``` ## ## will actually dump `x + y`, but at the same time will print at - ## compile time the expansion of the ``dump`` macro, which in this - ## case is ``debugEcho ["x + y", " = ", x + y]``. + ## compile time the expansion of the `dump` macro, which in this + ## case is `debugEcho ["x + y", " = ", x + y]`. echo body.toStrLit result = body +proc extractTypeImpl(n: NimNode): NimNode = + ## attempts to extract the type definition of the given symbol + case n.kind + of nnkSym: # can extract an impl + result = n.getImpl.extractTypeImpl() + of nnkObjectTy, nnkRefTy, nnkPtrTy: result = n + of nnkBracketExpr: + if n.typeKind == ntyTypeDesc: + result = n[1].extractTypeImpl() + else: + doAssert n.typeKind == ntyGenericInst + result = n[0].getImpl() + of nnkTypeDef: + result = n[2] + else: error("Invalid node to retrieve type implementation of: " & $n.kind) + proc customPragmaNode(n: NimNode): NimNode = - expectKind(n, {nnkSym, nnkDotExpr, nnkBracketExpr, nnkTypeOfExpr, nnkCheckedFieldExpr}) + expectKind(n, {nnkSym, nnkDotExpr, nnkBracketExpr, nnkTypeOfExpr, nnkType, nnkCheckedFieldExpr}) let typ = n.getTypeInst() if typ.kind == nnkBracketExpr and typ.len > 1 and typ[1].kind == nnkProcTy: return typ[1][1] elif typ.typeKind == ntyTypeDesc: - let impl = typ[1].getImpl() - if impl[0].kind == nnkPragmaExpr: + let impl = getImpl( + if kind(typ[1]) == nnkBracketExpr: typ[1][0] + else: typ[1] + ) + if impl.kind == nnkNilLit: + return impl + elif impl[0].kind == nnkPragmaExpr: return impl[0][1] else: return impl[0] # handle types which don't have macro at all @@ -1501,10 +1597,10 @@ proc customPragmaNode(n: NimNode): NimNode = let impl = n.getImpl() if impl.kind in RoutineNodes: return impl.pragma - elif impl.kind == nnkIdentDefs and impl[0].kind == nnkPragmaExpr: + elif impl.kind in {nnkIdentDefs, nnkConstDef} and impl[0].kind == nnkPragmaExpr: return impl[0][1] else: - let timpl = typ.getImpl() + let timpl = getImpl(if typ.kind == nnkBracketExpr: typ[0] else: typ) if timpl.len>0 and timpl[0].len>1: return timpl[0][1] else: @@ -1512,12 +1608,13 @@ proc customPragmaNode(n: NimNode): NimNode = if n.kind in {nnkDotExpr, nnkCheckedFieldExpr}: let name = $(if n.kind == nnkCheckedFieldExpr: n[0][1] else: n[1]) - let typInst = getTypeInst(if n.kind == nnkCheckedFieldExpr or n[0].kind == nnkHiddenDeref: n[0][0] else: n[0]) - var typDef = getImpl(if typInst.kind == nnkVarTy: typInst[0] else: typInst) + var typInst = getTypeInst(if n.kind == nnkCheckedFieldExpr or n[0].kind == nnkHiddenDeref: n[0][0] else: n[0]) + while typInst.kind in {nnkVarTy, nnkBracketExpr}: typInst = typInst[0] + var typDef = getImpl(typInst) while typDef != nil: typDef.expectKind(nnkTypeDef) - let typ = typDef[2] - typ.expectKind({nnkRefTy, nnkPtrTy, nnkObjectTy}) + let typ = typDef[2].extractTypeImpl() + if typ.kind notin {nnkRefTy, nnkPtrTy, nnkObjectTy}: break let isRef = typ.kind in {nnkRefTy, nnkPtrTy} if isRef and typ[0].kind in {nnkSym, nnkBracketExpr}: # defines ref type for another object(e.g. X = ref X) typDef = getImpl(typ[0]) @@ -1528,18 +1625,17 @@ proc customPragmaNode(n: NimNode): NimNode = for i in 0..<identDefsStack.len: identDefsStack[i] = obj[2][i] while identDefsStack.len > 0: var identDefs = identDefsStack.pop() - if identDefs.kind == nnkRecCase: - identDefsStack.add(identDefs[0]) - for i in 1..<identDefs.len: - let varNode = identDefs[i] - # if it is and empty branch, skip - if varNode[0].kind == nnkNilLit: continue - if varNode[1].kind == nnkIdentDefs: - identDefsStack.add(varNode[1]) - else: # nnkRecList - for j in 0 ..< varNode[1].len: - identDefsStack.add(varNode[1][j]) + case identDefs.kind + of nnkRecList: + for child in identDefs.children: + identDefsStack.add(child) + of nnkRecCase: + # Add condition definition + identDefsStack.add(identDefs[0]) + # Add branches + for i in 1 ..< identDefs.len: + identDefsStack.add(identDefs[i].last) else: for i in 0 .. identDefs.len - 3: let varNode = identDefs[i] @@ -1560,9 +1656,9 @@ macro hasCustomPragma*(n: typed, cp: typed{nkSym}): untyped = ## Expands to `true` if expression `n` which is expected to be `nnkDotExpr` ## (if checking a field), a proc or a type has custom pragma `cp`. ## - ## See also `getCustomPragmaVal`. + ## See also `getCustomPragmaVal`_. ## - ## .. code-block:: nim + ## ```nim ## template myAttr() {.pragma.} ## type ## MyObj = object @@ -1573,6 +1669,7 @@ macro hasCustomPragma*(n: typed, cp: typed{nkSym}): untyped = ## var o: MyObj ## assert(o.myField.hasCustomPragma(myAttr)) ## assert(myProc.hasCustomPragma(myAttr)) + ## ``` let pragmaNode = customPragmaNode(n) for p in pragmaNode: if (p.kind == nnkSym and p == cp) or @@ -1584,9 +1681,9 @@ macro getCustomPragmaVal*(n: typed, cp: typed{nkSym}): untyped = ## Expands to value of custom pragma `cp` of expression `n` which is expected ## to be `nnkDotExpr`, a proc or a type. ## - ## See also `hasCustomPragma` + ## See also `hasCustomPragma`_. ## - ## .. code-block:: nim + ## ```nim ## template serializationKey(key: string) {.pragma.} ## type ## MyObj {.serializationKey: "mo".} = object @@ -1595,11 +1692,12 @@ macro getCustomPragmaVal*(n: typed, cp: typed{nkSym}): untyped = ## assert(o.myField.getCustomPragmaVal(serializationKey) == "mf") ## assert(o.getCustomPragmaVal(serializationKey) == "mo") ## assert(MyObj.getCustomPragmaVal(serializationKey) == "mo") + ## ``` result = nil let pragmaNode = customPragmaNode(n) for p in pragmaNode: if p.kind in nnkPragmaCallKinds and p.len > 0 and p[0].kind == nnkSym and p[0] == cp: - if p.len == 2: + if p.len == 2 or (p.len == 3 and p[1].kind == nnkSym and p[1].symKind == nskType): result = p[1] else: let def = p[0].getImpl[3] @@ -1613,6 +1711,21 @@ macro getCustomPragmaVal*(n: typed, cp: typed{nkSym}): untyped = error(n.repr & " doesn't have a pragma named " & cp.repr()) # returning an empty node results in most cases in a cryptic error, macro unpackVarargs*(callee: untyped; args: varargs[untyped]): untyped = + ## Calls `callee` with `args` unpacked as individual arguments. + ## This is useful in 2 cases: + ## * when forwarding `varargs[T]` for some typed `T` + ## * when forwarding `varargs[untyped]` when `args` can potentially be empty, + ## due to a compiler limitation + runnableExamples: + template call1(fun: typed; args: varargs[untyped]): untyped = + unpackVarargs(fun, args) + # when varargsLen(args) > 0: fun(args) else: fun() # this would also work + template call2(fun: typed; args: varargs[typed]): untyped = + unpackVarargs(fun, args) + proc fn1(a = 0, b = 1) = discard (a, b) + call1(fn1, 10, 11) + call1(fn1) # `args` is empty in this case + if false: call2(echo, 10, 11) # would print 1011 result = newCall(callee) for i in 0 ..< args.len: result.add args[i] @@ -1624,17 +1737,17 @@ proc getProjectPath*(): string = discard ## which returns the path of the source file containing that template ## call. ## - ## For example, assume a ``dir1/foo.nim`` that imports a ``dir2/bar.nim``, - ## have the ``bar.nim`` print out both ``getProjectPath`` and - ## ``currentSourcePath`` outputs. + ## For example, assume a `dir1/foo.nim` that imports a `dir2/bar.nim`, + ## have the `bar.nim` print out both `getProjectPath` and + ## `currentSourcePath` outputs. ## - ## Now when ``foo.nim`` is compiled, the ``getProjectPath`` from - ## ``bar.nim`` will return the ``dir1/`` path, while the ``currentSourcePath`` - ## will return the path to the ``bar.nim`` source file. + ## Now when `foo.nim` is compiled, the `getProjectPath` from + ## `bar.nim` will return the `dir1/` path, while the `currentSourcePath` + ## will return the path to the `bar.nim` source file. ## - ## Now when ``bar.nim`` is compiled directly, the ``getProjectPath`` - ## will now return the ``dir2/`` path, and the ``currentSourcePath`` - ## will still return the same path, the path to the ``bar.nim`` source + ## Now when `bar.nim` is compiled directly, the `getProjectPath` + ## will now return the `dir2/` path, and the `currentSourcePath` + ## will still return the same path, the path to the `bar.nim` source ## file. ## ## The path returned by this proc is set at compile time. @@ -1642,22 +1755,21 @@ proc getProjectPath*(): string = discard ## See also: ## * `getCurrentDir proc <os.html#getCurrentDir>`_ -when defined(nimMacrosSizealignof): - proc getSize*(arg: NimNode): int {.magic: "NSizeOf", noSideEffect.} = - ## Returns the same result as ``system.sizeof`` if the size is - ## known by the Nim compiler. Returns a negative value if the Nim - ## compiler does not know the size. - proc getAlign*(arg: NimNode): int {.magic: "NSizeOf", noSideEffect.} = - ## Returns the same result as ``system.alignof`` if the alignment - ## is known by the Nim compiler. It works on ``NimNode`` for use - ## in macro context. Returns a negative value if the Nim compiler - ## does not know the alignment. - proc getOffset*(arg: NimNode): int {.magic: "NSizeOf", noSideEffect.} = - ## Returns the same result as ``system.offsetof`` if the offset is - ## known by the Nim compiler. It expects a resolved symbol node - ## from a field of a type. Therefore it only requires one argument - ## instead of two. Returns a negative value if the Nim compiler - ## does not know the offset. +proc getSize*(arg: NimNode): int {.magic: "NSizeOf", noSideEffect.} = + ## Returns the same result as `system.sizeof` if the size is + ## known by the Nim compiler. Returns a negative value if the Nim + ## compiler does not know the size. +proc getAlign*(arg: NimNode): int {.magic: "NSizeOf", noSideEffect.} = + ## Returns the same result as `system.alignof` if the alignment + ## is known by the Nim compiler. It works on `NimNode` for use + ## in macro context. Returns a negative value if the Nim compiler + ## does not know the alignment. +proc getOffset*(arg: NimNode): int {.magic: "NSizeOf", noSideEffect.} = + ## Returns the same result as `system.offsetof` if the offset is + ## known by the Nim compiler. It expects a resolved symbol node + ## from a field of a type. Therefore it only requires one argument + ## instead of two. Returns a negative value if the Nim compiler + ## does not know the offset. proc isExported*(n: NimNode): bool {.noSideEffect.} = ## Returns whether the symbol is exported or not. @@ -1667,33 +1779,34 @@ proc extractDocCommentsAndRunnables*(n: NimNode): NimNode = ## runnableExamples in `a`, stopping at the first child that is neither. ## Example: ## - ## .. code-block:: nim - ## import macros - ## macro transf(a): untyped = - ## result = quote do: - ## proc fun2*() = discard - ## let header = extractDocCommentsAndRunnables(a.body) - ## # correct usage: rest is appended - ## result.body = header - ## result.body.add quote do: discard # just an example - ## # incorrect usage: nesting inside a nnkStmtList: - ## # result.body = quote do: (`header`; discard) - ## - ## proc fun*() {.transf.} = - ## ## first comment - ## runnableExamples: discard - ## runnableExamples: discard - ## ## last comment - ## discard # first statement after doc comments + runnableExamples - ## ## not docgen'd + ## ```nim + ## import std/macros + ## macro transf(a): untyped = + ## result = quote do: + ## proc fun2*() = discard + ## let header = extractDocCommentsAndRunnables(a.body) + ## # correct usage: rest is appended + ## result.body = header + ## result.body.add quote do: discard # just an example + ## # incorrect usage: nesting inside a nnkStmtList: + ## # result.body = quote do: (`header`; discard) + ## + ## proc fun*() {.transf.} = + ## ## first comment + ## runnableExamples: discard + ## runnableExamples: discard + ## ## last comment + ## discard # first statement after doc comments + runnableExamples + ## ## not docgen'd + ## ``` result = newStmtList() for ni in n: case ni.kind of nnkCommentStmt: result.add ni - of nnkCall: - if ni[0].kind == nnkIdent and ni[0].strVal == "runnableExamples": + of nnkCall, nnkCommand: + if ni[0].kind == nnkIdent and ni[0].eqIdent "runnableExamples": result.add ni else: break else: break diff --git a/lib/core/rlocks.nim b/lib/core/rlocks.nim index 27a9c5e60..8cb0cef05 100644 --- a/lib/core/rlocks.nim +++ b/lib/core/rlocks.nim @@ -11,10 +11,12 @@ when not compileOption("threads") and not defined(nimdoc): - {.error: "Rlocks requires --threads:on option.".} + when false: + # make rlocks modlue consistent with locks module, + # so they can replace each other seamlessly. + {.error: "Rlocks requires --threads:on option.".} -const insideRLocksModule = true -include "system/syslocks" +import std/private/syslocks type RLock* = SysLock ## Nim lock, re-entrant @@ -29,27 +31,27 @@ proc initRLock*(lock: var RLock) {.inline.} = else: initSysLock(lock) -proc deinitRLock*(lock: var RLock) {.inline.} = +proc deinitRLock*(lock: RLock) {.inline.} = ## Frees the resources associated with the lock. deinitSys(lock) -proc tryAcquire*(lock: var RLock): bool = +proc tryAcquire*(lock: var RLock): bool {.inline.} = ## Tries to acquire the given lock. Returns `true` on success. result = tryAcquireSys(lock) -proc acquire*(lock: var RLock) = +proc acquire*(lock: var RLock) {.inline.} = ## Acquires the given lock. acquireSys(lock) -proc release*(lock: var RLock) = +proc release*(lock: var RLock) {.inline.} = ## Releases the given lock. releaseSys(lock) -template withRLock*(lock: var RLock, code: untyped): untyped = +template withRLock*(lock: RLock, code: untyped) = ## Acquires the given lock and then executes the code. - block: - acquire(lock) - defer: - release(lock) - {.locks: [lock].}: + acquire(lock) + {.locks: [lock].}: + try: code + finally: + release(lock) diff --git a/lib/core/typeinfo.nim b/lib/core/typeinfo.nim index c15f6dc1f..f2fee91c4 100644 --- a/lib/core/typeinfo.nim +++ b/lib/core/typeinfo.nim @@ -9,16 +9,29 @@ ## This module implements an interface to Nim's `runtime type information`:idx: ## (`RTTI`:idx:). See the `marshal <marshal.html>`_ module for an example of -## what this module allows you to do. +## what this allows you to do. ## -## Note that even though ``Any`` and its operations hide the nasty low level -## details from its clients, it remains inherently unsafe! Also, Nim's -## runtime type information will evolve and may eventually be deprecated. -## As an alternative approach to programmatically understanding and -## manipulating types, consider using the `macros <macros.html>`_ package to -## work with the types' AST representation at compile time. See, for example, -## the `getTypeImpl proc<macros.html#getTypeImpl,NimNode>`_. As an alternative -## approach to storing arbitrary types at runtime, consider using generics. +## .. note:: Even though `Any` and its operations hide the nasty low level +## details from its users, it remains inherently unsafe! Also, Nim's +## runtime type information will evolve and may eventually be deprecated. +## As an alternative approach to programmatically understanding and +## manipulating types, consider using the `macros <macros.html>`_ module to +## work with the types' AST representation at compile time. See for example +## the `getTypeImpl proc <macros.html#getTypeImpl,NimNode>`_. As an alternative +## approach to storing arbitrary types at runtime, consider using generics. + +runnableExamples: + var x: Any + + var i = 42 + x = i.toAny + assert x.kind == akInt + assert x.getInt == 42 + + var s = @[1, 2, 3] + x = s.toAny + assert x.kind == akSequence + assert x.len == 3 {.push hints: off.} @@ -27,45 +40,50 @@ include "system/hti.nim" {.pop.} +when defined(nimPreviewSlimSystem): + import std/assertions + + type - AnyKind* = enum ## what kind of ``any`` it is - akNone = 0, ## invalid any - akBool = 1, ## any represents a ``bool`` - akChar = 2, ## any represents a ``char`` - akEnum = 14, ## any represents an enum - akArray = 16, ## any represents an array - akObject = 17, ## any represents an object - akTuple = 18, ## any represents a tuple - akSet = 19, ## any represents a set - akRange = 20, ## any represents a range - akPtr = 21, ## any represents a ptr - akRef = 22, ## any represents a ref - akSequence = 24, ## any represents a sequence - akProc = 25, ## any represents a proc - akPointer = 26, ## any represents a pointer - akString = 28, ## any represents a string - akCString = 29, ## any represents a cstring - akInt = 31, ## any represents an int - akInt8 = 32, ## any represents an int8 - akInt16 = 33, ## any represents an int16 - akInt32 = 34, ## any represents an int32 - akInt64 = 35, ## any represents an int64 - akFloat = 36, ## any represents a float - akFloat32 = 37, ## any represents a float32 - akFloat64 = 38, ## any represents a float64 - akFloat128 = 39, ## any represents a float128 - akUInt = 40, ## any represents an unsigned int - akUInt8 = 41, ## any represents an unsigned int8 - akUInt16 = 42, ## any represents an unsigned in16 - akUInt32 = 43, ## any represents an unsigned int32 - akUInt64 = 44, ## any represents an unsigned int64 + AnyKind* = enum ## The kind of `Any`. + akNone = 0, ## invalid + akBool = 1, ## bool + akChar = 2, ## char + akEnum = 14, ## enum + akArray = 16, ## array + akObject = 17, ## object + akTuple = 18, ## tuple + akSet = 19, ## set + akRange = 20, ## range + akPtr = 21, ## ptr + akRef = 22, ## ref + akSequence = 24, ## sequence + akProc = 25, ## proc + akPointer = 26, ## pointer + akString = 28, ## string + akCString = 29, ## cstring + akInt = 31, ## int + akInt8 = 32, ## int8 + akInt16 = 33, ## int16 + akInt32 = 34, ## int32 + akInt64 = 35, ## int64 + akFloat = 36, ## float + akFloat32 = 37, ## float32 + akFloat64 = 38, ## float64 + akFloat128 = 39, ## float128 + akUInt = 40, ## uint + akUInt8 = 41, ## uint8 + akUInt16 = 42, ## uin16 + akUInt32 = 43, ## uint32 + akUInt64 = 44, ## uint64 # akOpt = 44+18 ## the builtin 'opt' type. - Any* = object ## can represent any nim value; NOTE: the wrapped - ## value can be modified with its wrapper! This means - ## that ``Any`` keeps a non-traced pointer to its - ## wrapped value and **must not** live longer than - ## its wrapped value. + Any* = object + ## A type that can represent any nim value. + ## + ## .. danger:: The wrapped value can be modified with its wrapper! This means + ## that `Any` keeps a non-traced pointer to its wrapped value and + ## **must not** live longer than its wrapped value. value: pointer when defined(js): rawType: PNimType @@ -73,7 +91,7 @@ type rawTypePtr: pointer ppointer = ptr pointer - pbyteArray = ptr array[0xffff, int8] + pbyteArray = ptr array[0xffff, uint8] when not defined(gcDestructors): type @@ -84,13 +102,15 @@ when not defined(gcDestructors): PGenSeq = ptr TGenericSeq when defined(gogc): - const GenericSeqSize = (3 * sizeof(int)) + const GenericSeqSize = 3 * sizeof(int) else: - const GenericSeqSize = (2 * sizeof(int)) + const GenericSeqSize = 2 * sizeof(int) else: include system/seqs_v2_reimpl +from std/private/strimpl import cmpNimIdentifier + when not defined(js): template rawType(x: Any): PNimType = cast[PNimType](x.rawTypePtr) @@ -101,8 +121,7 @@ when not defined(js): proc genericAssign(dest, src: pointer, mt: PNimType) {.importCompilerProc.} when not defined(gcDestructors): - proc genericShallowAssign(dest, src: pointer, mt: PNimType) {. - importCompilerProc.} + proc genericShallowAssign(dest, src: pointer, mt: PNimType) {.importCompilerProc.} proc incrSeq(seq: PGenSeq, elemSize, elemAlign: int): PGenSeq {.importCompilerProc.} proc newObj(typ: PNimType, size: int): pointer {.importCompilerProc.} proc newSeq(typ: PNimType, len: int): pointer {.importCompilerProc.} @@ -110,29 +129,31 @@ when not defined(gcDestructors): else: proc nimNewObj(size, align: int): pointer {.importCompilerProc.} proc newSeqPayload(cap, elemSize, elemAlign: int): pointer {.importCompilerProc.} - proc prepareSeqAdd(len: int; p: pointer; addlen, elemSize, elemAlign: int): pointer {. + proc prepareSeqAddUninit(len: int; p: pointer; addlen, elemSize, elemAlign: int): pointer {. + importCompilerProc.} + proc zeroNewElements(len: int; p: pointer; addlen, elemSize, elemAlign: int) {. importCompilerProc.} -template `+!!`(a, b): untyped = cast[pointer](cast[ByteAddress](a) + b) +template `+!!`(a, b): untyped = cast[pointer](cast[int](a) + b) proc getDiscriminant(aa: pointer, n: ptr TNimNode): int = assert(n.kind == nkCase) var d: int - var a = cast[ByteAddress](aa) + let a = cast[int](aa) case n.typ.size - of 1: d = ze(cast[ptr int8](a +% n.offset)[]) - of 2: d = ze(cast[ptr int16](a +% n.offset)[]) - of 4: d = int(cast[ptr int32](a +% n.offset)[]) - of 8: d = int(cast[ptr int64](a +% n.offset)[]) + of 1: d = int(cast[ptr uint8](a +% n.offset)[]) + of 2: d = int(cast[ptr uint16](a +% n.offset)[]) + of 4: d = int(cast[ptr uint32](a +% n.offset)[]) + of 8: d = int(cast[ptr uint64](a +% n.offset)[]) else: assert(false) return d proc selectBranch(aa: pointer, n: ptr TNimNode): ptr TNimNode = - var discr = getDiscriminant(aa, n) + let discr = getDiscriminant(aa, n) if discr <% n.len: result = n.sons[discr] if result == nil: result = n.sons[n.len] - # n.sons[n.len] contains the ``else`` part (but may be nil) + # n.sons[n.len] contains the `else` part (but may be nil) else: result = n.sons[n.len] @@ -140,19 +161,9 @@ proc newAny(value: pointer, rawType: PNimType): Any {.inline.} = result.value = value result.rawType = rawType -when declared(system.VarSlot): - proc toAny*(x: VarSlot): Any {.inline.} = - ## Constructs a ``Any`` object from a variable slot ``x``. - ## This captures `x`'s address, so `x` can be modified with its - ## ``Any`` wrapper! The client needs to ensure that the wrapper - ## **does not** live longer than `x`! - ## This is provided for easier reflection capabilities of a debugger. - result.value = x.address - result.rawType = x.typ - proc toAny*[T](x: var T): Any {.inline.} = - ## Constructs a ``Any`` object from `x`. This captures `x`'s address, so - ## `x` can be modified with its ``Any`` wrapper! The client needs to ensure + ## Constructs an `Any` object from `x`. This captures `x`'s address, so + ## `x` can be modified with its `Any` wrapper! The caller needs to ensure ## that the wrapper **does not** live longer than `x`! newAny(addr(x), cast[PNimType](getTypeInfo(x))) @@ -165,17 +176,17 @@ proc size*(x: Any): int {.inline.} = result = x.rawType.size proc baseTypeKind*(x: Any): AnyKind {.inline.} = - ## Gets the base type's kind; ``akNone`` is returned if `x` has no base type. + ## Gets the base type's kind. If `x` has no base type, `akNone` is returned. if x.rawType.base != nil: result = AnyKind(ord(x.rawType.base.kind)) proc baseTypeSize*(x: Any): int {.inline.} = - ## Returns the size of `x`'s basetype. + ## Returns the size of `x`'s base type. If `x` has no base type, 0 is returned. if x.rawType.base != nil: result = x.rawType.base.size proc invokeNew*(x: Any) = - ## Performs ``new(x)``. `x` needs to represent a ``ref``. + ## Performs `new(x)`. `x` needs to represent a `ref`. assert x.rawType.kind == tyRef when defined(gcDestructors): cast[ppointer](x.value)[] = nimNewObj(x.rawType.base.size, x.rawType.base.align) @@ -184,7 +195,7 @@ proc invokeNew*(x: Any) = genericAssign(x.value, addr(z), x.rawType) proc invokeNewSeq*(x: Any, len: int) = - ## Performs ``newSeq(x, len)``. `x` needs to represent a ``seq``. + ## Performs `newSeq(x, len)`. `x` needs to represent a `seq`. assert x.rawType.kind == tySequence when defined(gcDestructors): var s = cast[ptr NimSeqV2Reimpl](x.value) @@ -196,12 +207,14 @@ proc invokeNewSeq*(x: Any, len: int) = genericShallowAssign(x.value, addr(z), x.rawType) proc extendSeq*(x: Any) = - ## Performs ``setLen(x, x.len+1)``. `x` needs to represent a ``seq``. + ## Performs `setLen(x, x.len+1)`. `x` needs to represent a `seq`. assert x.rawType.kind == tySequence when defined(gcDestructors): var s = cast[ptr NimSeqV2Reimpl](x.value) let elem = x.rawType.base - s.p = cast[ptr NimSeqPayloadReimpl](prepareSeqAdd(s.len, s.p, 1, elem.size, elem.align)) + if s.p == nil or s.p.cap < s.len+1: + s.p = cast[ptr NimSeqPayloadReimpl](prepareSeqAddUninit(s.len, s.p, 1, elem.size, elem.align)) + zeroNewElements(s.len, s.p, 1, elem.size, elem.align) inc s.len else: var y = cast[ptr PGenSeq](x.value)[] @@ -299,33 +312,32 @@ proc len*(x: Any): int = proc base*(x: Any): Any = - ## Returns base Any (useful for inherited object types). + ## Returns the base type of `x` (useful for inherited object types). result.rawType = x.rawType.base result.value = x.value proc isNil*(x: Any): bool = - ## `isNil` for an any `x` that represents a cstring, proc or + ## `isNil` for an `x` that represents a cstring, proc or ## some pointer type. - assert x.rawType.kind in {tyCString, tyRef, tyPtr, tyPointer, tyProc} + assert x.rawType.kind in {tyCstring, tyRef, tyPtr, tyPointer, tyProc} result = isNil(cast[ppointer](x.value)[]) -const - pointerLike = when defined(gcDestructors): {tyCString, tyRef, tyPtr, tyPointer, tyProc} - else: {tyString, tyCString, tyRef, tyPtr, tyPointer, - tySequence, tyProc} +const pointerLike = + when defined(gcDestructors): {tyCstring, tyRef, tyPtr, tyPointer, tyProc} + else: {tyString, tyCstring, tyRef, tyPtr, tyPointer, tySequence, tyProc} proc getPointer*(x: Any): pointer = - ## Retrieves the pointer value out of `x`. ``x`` needs to be of kind - ## ``akString``, ``akCString``, ``akProc``, ``akRef``, ``akPtr``, - ## ``akPointer``, ``akSequence``. + ## Retrieves the pointer value out of `x`. `x` needs to be of kind + ## `akString`, `akCString`, `akProc`, `akRef`, `akPtr`, + ## `akPointer` or `akSequence`. assert x.rawType.kind in pointerLike result = cast[ppointer](x.value)[] proc setPointer*(x: Any, y: pointer) = - ## Sets the pointer value of `x`. ``x`` needs to be of kind - ## ``akString``, ``akCString``, ``akProc``, ``akRef``, ``akPtr``, - ## ``akPointer``, ``akSequence``. + ## Sets the pointer value of `x`. `x` needs to be of kind + ## `akString`, `akCString`, `akProc`, `akRef`, `akPtr`, + ## `akPointer` or `akSequence`. assert x.rawType.kind in pointerLike if y != nil and x.rawType.kind != tyPointer: genericAssign(x.value, y, x.rawType) @@ -347,15 +359,15 @@ proc fieldsAux(p: pointer, n: ptr TNimNode, if m != nil: fieldsAux(p, m, ret) iterator fields*(x: Any): tuple[name: string, any: Any] = - ## Iterates over every active field of the any `x` that represents an object + ## Iterates over every active field of `x`. `x` needs to represent an object ## or a tuple. assert x.rawType.kind in {tyTuple, tyObject} - var p = x.value + let p = x.value var t = x.rawType # XXX BUG: does not work yet, however is questionable anyway when false: if x.rawType.kind == tyObject: t = cast[ptr PNimType](x.value)[] - var ret: seq[tuple[name: cstring, any: Any]] = @[] + var ret: seq[tuple[name: cstring, any: Any]] if t.kind == tyObject: while true: fieldsAux(p, t.node, ret) @@ -366,49 +378,31 @@ iterator fields*(x: Any): tuple[name: string, any: Any] = for name, any in items(ret): yield ($name, any) -proc cmpIgnoreStyle(a, b: cstring): int {.noSideEffect.} = - proc toLower(c: char): char {.inline.} = - if c in {'A'..'Z'}: result = chr(ord(c) + (ord('a') - ord('A'))) - else: result = c - var i = 0 - var j = 0 - if a[0] != b[0]: return 1 - while true: - while a[i] == '_': inc(i) - while b[j] == '_': inc(j) # BUGFIX: typo - var aa = toLower(a[i]) - var bb = toLower(b[j]) - result = ord(aa) - ord(bb) - if result != 0 or aa == '\0': break - inc(i) - inc(j) - -proc getFieldNode(p: pointer, n: ptr TNimNode, - name: cstring): ptr TNimNode = +proc getFieldNode(p: pointer, n: ptr TNimNode, name: cstring): ptr TNimNode = case n.kind of nkNone: assert(false) of nkSlot: - if cmpIgnoreStyle(n.name, name) == 0: + if cmpNimIdentifier(n.name, name) == 0: result = n of nkList: for i in 0..n.len-1: result = getFieldNode(p, n.sons[i], name) if result != nil: break of nkCase: - if cmpIgnoreStyle(n.name, name) == 0: + if cmpNimIdentifier(n.name, name) == 0: result = n else: - var m = selectBranch(p, n) + let m = selectBranch(p, n) if m != nil: result = getFieldNode(p, m, name) proc `[]=`*(x: Any, fieldName: string, value: Any) = - ## Sets a field of `x`; `x` represents an object or a tuple. + ## Sets a field of `x`. `x` needs to represent an object or a tuple. var t = x.rawType # XXX BUG: does not work yet, however is questionable anyway when false: if x.rawType.kind == tyObject: t = cast[ptr PNimType](x.value)[] assert x.rawType.kind in {tyTuple, tyObject} - var n = getFieldNode(x.value, t.node, fieldName) + let n = getFieldNode(x.value, t.node, fieldName) if n != nil: assert n.typ == value.rawType genericAssign(x.value +!! n.offset, value.value, value.rawType) @@ -416,13 +410,13 @@ proc `[]=`*(x: Any, fieldName: string, value: Any) = raise newException(ValueError, "invalid field name: " & fieldName) proc `[]`*(x: Any, fieldName: string): Any = - ## Gets a field of `x`; `x` represents an object or a tuple. + ## Gets a field of `x`. `x` needs to represent an object or a tuple. var t = x.rawType # XXX BUG: does not work yet, however is questionable anyway when false: if x.rawType.kind == tyObject: t = cast[ptr PNimType](x.value)[] assert x.rawType.kind in {tyTuple, tyObject} - var n = getFieldNode(x.value, t.node, fieldName) + let n = getFieldNode(x.value, t.node, fieldName) if n != nil: result.value = x.value +!! n.offset result.rawType = n.typ @@ -432,47 +426,47 @@ proc `[]`*(x: Any, fieldName: string): Any = raise newException(ValueError, "invalid field name: " & fieldName) proc `[]`*(x: Any): Any = - ## Dereference operation for the any `x` that represents a ptr or a ref. + ## Dereference operator for `Any`. `x` needs to represent a ptr or a ref. assert x.rawType.kind in {tyRef, tyPtr} result.value = cast[ppointer](x.value)[] result.rawType = x.rawType.base proc `[]=`*(x, y: Any) = - ## Dereference operation for the any `x` that represents a ptr or a ref. + ## Dereference operator for `Any`. `x` needs to represent a ptr or a ref. assert x.rawType.kind in {tyRef, tyPtr} assert y.rawType == x.rawType.base genericAssign(cast[ppointer](x.value)[], y.value, y.rawType) proc getInt*(x: Any): int = - ## Retrieves the int value out of `x`. `x` needs to represent an int. + ## Retrieves the `int` value out of `x`. `x` needs to represent an `int`. assert skipRange(x.rawType).kind == tyInt result = cast[ptr int](x.value)[] proc getInt8*(x: Any): int8 = - ## Retrieves the int8 value out of `x`. `x` needs to represent an int8. + ## Retrieves the `int8` value out of `x`. `x` needs to represent an `int8`. assert skipRange(x.rawType).kind == tyInt8 result = cast[ptr int8](x.value)[] proc getInt16*(x: Any): int16 = - ## Retrieves the int16 value out of `x`. `x` needs to represent an int16. + ## Retrieves the `int16` value out of `x`. `x` needs to represent an `int16`. assert skipRange(x.rawType).kind == tyInt16 result = cast[ptr int16](x.value)[] proc getInt32*(x: Any): int32 = - ## Retrieves the int32 value out of `x`. `x` needs to represent an int32. + ## Retrieves the `int32` value out of `x`. `x` needs to represent an `int32`. assert skipRange(x.rawType).kind == tyInt32 result = cast[ptr int32](x.value)[] proc getInt64*(x: Any): int64 = - ## Retrieves the int64 value out of `x`. `x` needs to represent an int64. + ## Retrieves the `int64` value out of `x`. `x` needs to represent an `int64`. assert skipRange(x.rawType).kind == tyInt64 result = cast[ptr int64](x.value)[] proc getBiggestInt*(x: Any): BiggestInt = ## Retrieves the integer value out of `x`. `x` needs to represent ## some integer, a bool, a char, an enum or a small enough bit set. - ## The value might be sign-extended to ``BiggestInt``. - var t = skipRange(x.rawType) + ## The value might be sign-extended to `BiggestInt`. + let t = skipRange(x.rawType) case t.kind of tyInt: result = BiggestInt(cast[ptr int](x.value)[]) of tyInt8: result = BiggestInt(cast[ptr int8](x.value)[]) @@ -483,8 +477,8 @@ proc getBiggestInt*(x: Any): BiggestInt = of tyChar: result = BiggestInt(cast[ptr char](x.value)[]) of tyEnum, tySet: case t.size - of 1: result = ze64(cast[ptr int8](x.value)[]) - of 2: result = ze64(cast[ptr int16](x.value)[]) + of 1: result = int64(cast[ptr uint8](x.value)[]) + of 2: result = int64(cast[ptr uint16](x.value)[]) of 4: result = BiggestInt(cast[ptr int32](x.value)[]) of 8: result = BiggestInt(cast[ptr int64](x.value)[]) else: assert false @@ -497,7 +491,7 @@ proc getBiggestInt*(x: Any): BiggestInt = proc setBiggestInt*(x: Any, y: BiggestInt) = ## Sets the integer value of `x`. `x` needs to represent ## some integer, a bool, a char, an enum or a small enough bit set. - var t = skipRange(x.rawType) + let t = skipRange(x.rawType) case t.kind of tyInt: cast[ptr int](x.value)[] = int(y) of tyInt8: cast[ptr int8](x.value)[] = int8(y) @@ -508,8 +502,8 @@ proc setBiggestInt*(x: Any, y: BiggestInt) = of tyChar: cast[ptr char](x.value)[] = chr(y.int) of tyEnum, tySet: case t.size - of 1: cast[ptr int8](x.value)[] = toU8(y.int) - of 2: cast[ptr int16](x.value)[] = toU16(y.int) + of 1: cast[ptr uint8](x.value)[] = uint8(y.int) + of 2: cast[ptr uint16](x.value)[] = uint16(y.int) of 4: cast[ptr int32](x.value)[] = int32(y) of 8: cast[ptr int64](x.value)[] = y else: assert false @@ -520,38 +514,34 @@ proc setBiggestInt*(x: Any, y: BiggestInt) = else: assert false proc getUInt*(x: Any): uint = - ## Retrieves the uint value out of `x`, `x` needs to represent an uint. + ## Retrieves the `uint` value out of `x`. `x` needs to represent a `uint`. assert skipRange(x.rawType).kind == tyUInt result = cast[ptr uint](x.value)[] proc getUInt8*(x: Any): uint8 = - ## Retrieves the uint8 value out of `x`, `x` needs to represent an - ## uint8. + ## Retrieves the `uint8` value out of `x`. `x` needs to represent a `uint8`. assert skipRange(x.rawType).kind == tyUInt8 result = cast[ptr uint8](x.value)[] proc getUInt16*(x: Any): uint16 = - ## Retrieves the uint16 value out of `x`, `x` needs to represent an - ## uint16. + ## Retrieves the `uint16` value out of `x`. `x` needs to represent a `uint16`. assert skipRange(x.rawType).kind == tyUInt16 result = cast[ptr uint16](x.value)[] proc getUInt32*(x: Any): uint32 = - ## Retrieves the uint32 value out of `x`, `x` needs to represent an - ## uint32. + ## Retrieves the `uint32` value out of `x`. `x` needs to represent a `uint32`. assert skipRange(x.rawType).kind == tyUInt32 result = cast[ptr uint32](x.value)[] proc getUInt64*(x: Any): uint64 = - ## Retrieves the uint64 value out of `x`, `x` needs to represent an - ## uint64. + ## Retrieves the `uint64` value out of `x`. `x` needs to represent a `uint64`. assert skipRange(x.rawType).kind == tyUInt64 result = cast[ptr uint64](x.value)[] proc getBiggestUint*(x: Any): uint64 = ## Retrieves the unsigned integer value out of `x`. `x` needs to ## represent an unsigned integer. - var t = skipRange(x.rawType) + let t = skipRange(x.rawType) case t.kind of tyUInt: result = uint64(cast[ptr uint](x.value)[]) of tyUInt8: result = uint64(cast[ptr uint8](x.value)[]) @@ -561,9 +551,9 @@ proc getBiggestUint*(x: Any): uint64 = else: assert false proc setBiggestUint*(x: Any; y: uint64) = - ## Sets the unsigned integer value of `c`. `c` needs to represent an + ## Sets the unsigned integer value of `x`. `x` needs to represent an ## unsigned integer. - var t = skipRange(x.rawType) + let t = skipRange(x.rawType) case t.kind: of tyUInt: cast[ptr uint](x.value)[] = uint(y) of tyUInt8: cast[ptr uint8](x.value)[] = uint8(y) @@ -573,14 +563,14 @@ proc setBiggestUint*(x: Any; y: uint64) = else: assert false proc getChar*(x: Any): char = - ## Retrieves the char value out of `x`. `x` needs to represent a char. - var t = skipRange(x.rawType) + ## Retrieves the `char` value out of `x`. `x` needs to represent a `char`. + let t = skipRange(x.rawType) assert t.kind == tyChar result = cast[ptr char](x.value)[] proc getBool*(x: Any): bool = - ## Retrieves the bool value out of `x`. `x` needs to represent a bool. - var t = skipRange(x.rawType) + ## Retrieves the `bool` value out of `x`. `x` needs to represent a `bool`. + let t = skipRange(x.rawType) assert t.kind == tyBool result = cast[ptr bool](x.value)[] @@ -593,13 +583,13 @@ proc skipRange*(x: Any): Any = proc getEnumOrdinal*(x: Any, name: string): int = ## Gets the enum field ordinal from `name`. `x` needs to represent an enum ## but is only used to access the type information. In case of an error - ## ``low(int)`` is returned. - var typ = skipRange(x.rawType) + ## `low(int)` is returned. + let typ = skipRange(x.rawType) assert typ.kind == tyEnum - var n = typ.node - var s = n.sons + let n = typ.node + let s = n.sons for i in 0 .. n.len-1: - if cmpIgnoreStyle($s[i].name, name) == 0: + if cmpNimIdentifier($s[i].name, name) == 0: if ntfEnumHole notin typ.flags: return i else: @@ -610,16 +600,16 @@ proc getEnumField*(x: Any, ordinalValue: int): string = ## Gets the enum field name as a string. `x` needs to represent an enum ## but is only used to access the type information. The field name of ## `ordinalValue` is returned. - var typ = skipRange(x.rawType) + let typ = skipRange(x.rawType) assert typ.kind == tyEnum - var e = ordinalValue + let e = ordinalValue if ntfEnumHole notin typ.flags: if e <% typ.node.len: return $typ.node.sons[e].name else: # ugh we need a slow linear search: - var n = typ.node - var s = n.sons + let n = typ.node + let s = n.sons for i in 0 .. n.len-1: if s[i].offset == e: return $s[i].name result = $e @@ -629,23 +619,23 @@ proc getEnumField*(x: Any): string = result = getEnumField(x, getBiggestInt(x).int) proc getFloat*(x: Any): float = - ## Retrieves the float value out of `x`. `x` needs to represent an float. + ## Retrieves the `float` value out of `x`. `x` needs to represent a `float`. assert skipRange(x.rawType).kind == tyFloat result = cast[ptr float](x.value)[] proc getFloat32*(x: Any): float32 = - ## Retrieves the float32 value out of `x`. `x` needs to represent an float32. + ## Retrieves the `float32` value out of `x`. `x` needs to represent a `float32`. assert skipRange(x.rawType).kind == tyFloat32 result = cast[ptr float32](x.value)[] proc getFloat64*(x: Any): float64 = - ## Retrieves the float64 value out of `x`. `x` needs to represent an float64. + ## Retrieves the `float64` value out of `x`. `x` needs to represent a `float64`. assert skipRange(x.rawType).kind == tyFloat64 result = cast[ptr float64](x.value)[] proc getBiggestFloat*(x: Any): BiggestFloat = ## Retrieves the float value out of `x`. `x` needs to represent - ## some float. The value is extended to ``BiggestFloat``. + ## some float. The value is extended to `BiggestFloat`. case skipRange(x.rawType).kind of tyFloat: result = BiggestFloat(cast[ptr float](x.value)[]) of tyFloat32: result = BiggestFloat(cast[ptr float32](x.value)[]) @@ -662,7 +652,7 @@ proc setBiggestFloat*(x: Any, y: BiggestFloat) = else: assert false proc getString*(x: Any): string = - ## Retrieves the string value out of `x`. `x` needs to represent a string. + ## Retrieves the `string` value out of `x`. `x` needs to represent a `string`. assert x.rawType.kind == tyString when defined(gcDestructors): result = cast[ptr string](x.value)[] @@ -671,50 +661,50 @@ proc getString*(x: Any): string = result = cast[ptr string](x.value)[] proc setString*(x: Any, y: string) = - ## Sets the string value of `x`. `x` needs to represent a string. + ## Sets the `string` value of `x`. `x` needs to represent a `string`. assert x.rawType.kind == tyString cast[ptr string](x.value)[] = y # also correct for gcDestructors proc getCString*(x: Any): cstring = - ## Retrieves the cstring value out of `x`. `x` needs to represent a cstring. - assert x.rawType.kind == tyCString + ## Retrieves the `cstring` value out of `x`. `x` needs to represent a `cstring`. + assert x.rawType.kind == tyCstring result = cast[ptr cstring](x.value)[] proc assign*(x, y: Any) = - ## Copies the value of `y` to `x`. The assignment operator for ``Any`` + ## Copies the value of `y` to `x`. The assignment operator for `Any` ## does NOT do this; it performs a shallow copy instead! assert y.rawType == x.rawType genericAssign(x.value, y.value, y.rawType) iterator elements*(x: Any): int = - ## Iterates over every element of `x` that represents a Nim bitset. + ## Iterates over every element of `x`. `x` needs to represent a `set`. assert x.rawType.kind == tySet - var typ = x.rawType - var p = x.value + let typ = x.rawType + let p = x.value # "typ.slots.len" field is for sets the "first" field var u: int64 case typ.size - of 1: u = ze64(cast[ptr int8](p)[]) - of 2: u = ze64(cast[ptr int16](p)[]) - of 4: u = ze64(cast[ptr int32](p)[]) + of 1: u = int64(cast[ptr uint8](p)[]) + of 2: u = int64(cast[ptr uint16](p)[]) + of 4: u = int64(cast[ptr uint32](p)[]) of 8: u = cast[ptr int64](p)[] else: - var a = cast[pbyteArray](p) + let a = cast[pbyteArray](p) for i in 0 .. typ.size*8-1: - if (ze(a[i div 8]) and (1 shl (i mod 8))) != 0: - yield i+typ.node.len + if (int(a[i div 8]) and (1 shl (i mod 8))) != 0: + yield i + typ.node.len if typ.size <= 8: for i in 0..sizeof(int64)*8-1: if (u and (1'i64 shl int64(i))) != 0'i64: - yield i+typ.node.len + yield i + typ.node.len proc inclSetElement*(x: Any, elem: int) = ## Includes an element `elem` in `x`. `x` needs to represent a Nim bitset. assert x.rawType.kind == tySet - var typ = x.rawType - var p = x.value + let typ = x.rawType + let p = x.value # "typ.slots.len" field is for sets the "first" field - var e = elem - typ.node.len + let e = elem - typ.node.len case typ.size of 1: var a = cast[ptr int8](p) @@ -730,4 +720,4 @@ proc inclSetElement*(x: Any, elem: int) = a[] = a[] or (1'i64 shl e) else: var a = cast[pbyteArray](p) - a[e shr 3] = toU8(a[e shr 3] or (1 shl (e and 7))) + a[e shr 3] = a[e shr 3] or uint8(1 shl (e and 7)) |