mirror of
https://github.com/nim-lang/Nim.git
synced 2026-02-17 16:38:33 +00:00
Merge branch 'devel' into araq-big-refactoring
This commit is contained in:
@@ -56,6 +56,7 @@
|
||||
- Added the type ``times.Duration`` for representing fixed durations of time.
|
||||
- Added the proc ``times.convert`` for converting between different time units,
|
||||
e.g days to seconds.
|
||||
- Added the proc ``algorithm.binarySearch[T, K]`` with the ```cmp``` parameter.
|
||||
|
||||
### Library changes
|
||||
|
||||
@@ -74,6 +75,7 @@
|
||||
deprecated.
|
||||
- The `terminal` module now exports additional procs for generating ANSI color
|
||||
codes as strings.
|
||||
- Added the parameter ``val`` for the ``CritBitTree[int].inc`` proc.
|
||||
|
||||
### Language additions
|
||||
|
||||
|
||||
@@ -1116,8 +1116,8 @@ proc registerModuleToMain(g: BModuleList; m: PSym) =
|
||||
var
|
||||
init = m.getInitName
|
||||
datInit = m.getDatInitName
|
||||
addf(g.mainModProcs, "NIM_EXTERNC N_NOINLINE(void, $1)(void);$N", [init])
|
||||
addf(g.mainModProcs, "NIM_EXTERNC N_NOINLINE(void, $1)(void);$N", [datInit])
|
||||
addf(g.mainModProcs, "N_LIB_PRIVATE N_NIMCALL(void, $1)(void);$N", [init])
|
||||
addf(g.mainModProcs, "N_LIB_PRIVATE N_NIMCALL(void, $1)(void);$N", [datInit])
|
||||
if sfSystemModule notin m.flags:
|
||||
addf(g.mainDatInit, "\t$1();$N", [datInit])
|
||||
let initCall = "\t$1();$N" % [init]
|
||||
@@ -1128,7 +1128,7 @@ proc registerModuleToMain(g: BModuleList; m: PSym) =
|
||||
|
||||
proc genInitCode(m: BModule) =
|
||||
var initname = getInitName(m.module)
|
||||
var prc = "NIM_EXTERNC N_NOINLINE(void, $1)(void) {$N" % [initname]
|
||||
var prc = "N_LIB_PRIVATE N_NIMCALL(void, $1)(void) {$N" % [initname]
|
||||
if m.typeNodes > 0:
|
||||
appcg(m, m.s[cfsTypeInit1], "static #TNimNode $1[$2];$n",
|
||||
[m.typeNodesName, rope(m.typeNodes)])
|
||||
@@ -1170,7 +1170,7 @@ proc genInitCode(m: BModule) =
|
||||
add(prc, deinitGCFrame(m.initProc))
|
||||
addf(prc, "}$N$N", [])
|
||||
|
||||
prc.addf("NIM_EXTERNC N_NOINLINE(void, $1)(void) {$N",
|
||||
prc.addf("N_LIB_PRIVATE N_NIMCALL(void, $1)(void) {$N",
|
||||
[getDatInitName(m.module)])
|
||||
|
||||
for i in cfsTypeInit1..cfsDynLibInit:
|
||||
|
||||
@@ -189,7 +189,7 @@ proc putComment(g: var TSrcGen, s: string) =
|
||||
put(g, tkComment, com)
|
||||
com = "## "
|
||||
inc(i)
|
||||
if s[i] == '\x0A': inc(i)
|
||||
if i < s.len and s[i] == '\x0A': inc(i)
|
||||
optNL(g, ind)
|
||||
of '\x0A':
|
||||
put(g, tkComment, com)
|
||||
|
||||
@@ -441,7 +441,7 @@ proc setLenSeq(c: PCtx; node: PNode; newLen: int; info: TLineInfo) =
|
||||
node.sons[i] = newNodeI(typeKind, info)
|
||||
|
||||
const
|
||||
errIndexOutOfBounds = "index ouf of bounds"
|
||||
errIndexOutOfBounds = "index out of bounds"
|
||||
errNilAccess = "attempt to access a nil address"
|
||||
errOverOrUnderflow = "over- or underflow"
|
||||
errConstantDivisionByZero = "division by zero"
|
||||
|
||||
@@ -243,3 +243,11 @@ vcc.cpp.options.size = "/O1"
|
||||
|
||||
# Configuration for the Tiny C Compiler:
|
||||
tcc.options.always = "-w"
|
||||
|
||||
# Configuration for the Genode toolchain
|
||||
amd64.genode.gcc.cpp.exe = "genode-x86-g++"
|
||||
amd64.genode.gcc.exe = "genode-x86-gcc"
|
||||
amd64.genode.gcc.path = "/usr/local/genode-gcc/bin"
|
||||
arm.genode.gcc.cpp.exe = "genode-arm-g++"
|
||||
arm.genode.gcc.exe = "genode-arm-gcc"
|
||||
arm.genode.gcc.path = "/usr/local/genode-gcc/bin"
|
||||
|
||||
@@ -128,7 +128,7 @@ The ``watchpoint`` pragma is syntactically a statement. It can be used
|
||||
to mark a location as a watchpoint:
|
||||
|
||||
.. code-block:: Nim
|
||||
var a: array [0..20, int]
|
||||
var a: array[0..20, int]
|
||||
|
||||
{.watchpoint: a[3].}
|
||||
for i in 0 .. 20: a[i] = i
|
||||
|
||||
@@ -2527,7 +2527,7 @@ The implicit initialization can be avoided for optimization reasons with the
|
||||
|
||||
.. code-block:: nim
|
||||
var
|
||||
a {.noInit.}: array [0..1023, char]
|
||||
a {.noInit.}: array[0..1023, char]
|
||||
|
||||
If a proc is annotated with the ``noinit`` pragma this refers to its implicit
|
||||
``result`` variable:
|
||||
@@ -6632,7 +6632,7 @@ Syntactically it has to be used as a statement inside the loop:
|
||||
enumA, enumB, enumC, enumD, enumE
|
||||
|
||||
proc vm() =
|
||||
var instructions: array [0..100, MyEnum]
|
||||
var instructions: array[0..100, MyEnum]
|
||||
instructions[2] = enumC
|
||||
instructions[3] = enumD
|
||||
instructions[4] = enumA
|
||||
|
||||
@@ -64,27 +64,49 @@ proc reversed*[T](a: openArray[T]): seq[T] =
|
||||
## returns the reverse of the array `a`.
|
||||
reversed(a, 0, a.high)
|
||||
|
||||
proc binarySearch*[T, K](a: openArray[T], key: K,
|
||||
cmp: proc (x: T, y: K): int {.closure.}): int =
|
||||
## binary search for `key` in `a`. Returns -1 if not found.
|
||||
##
|
||||
## `cmp` is the comparator function to use, the expected return values are
|
||||
## the same as that of system.cmp.
|
||||
if a.len == 0:
|
||||
return -1
|
||||
|
||||
let len = a.len
|
||||
|
||||
if len == 1:
|
||||
if cmp(a[0], key) == 0:
|
||||
return 0
|
||||
else:
|
||||
return -1
|
||||
|
||||
if (len and (len - 1)) == 0:
|
||||
# when `len` is a power of 2, a faster shr can be used.
|
||||
var step = len shr 1
|
||||
while step > 0:
|
||||
let i = result or step
|
||||
if cmp(a[i], key) < 1:
|
||||
result = i
|
||||
step = step shr 1
|
||||
if cmp(a[result], key) != 0: result = -1
|
||||
else:
|
||||
var b = len
|
||||
while result < b:
|
||||
var mid = (result + b) shr 1
|
||||
if cmp(a[mid], key) < 0:
|
||||
result = mid + 1
|
||||
else:
|
||||
b = mid
|
||||
if result >= len or cmp(a[result], key) != 0: result = -1
|
||||
|
||||
proc binarySearch*[T](a: openArray[T], key: T): int =
|
||||
## binary search for `key` in `a`. Returns -1 if not found.
|
||||
if ((a.len - 1) and a.len) == 0 and a.len > 0:
|
||||
# when `a.len` is a power of 2, a faster div can be used.
|
||||
var step = a.len div 2
|
||||
while step > 0:
|
||||
if a[result or step] <= key:
|
||||
result = result or step
|
||||
step = step shr 1
|
||||
if a[result] != key: result = -1
|
||||
else:
|
||||
var b = len(a)
|
||||
while result < b:
|
||||
var mid = (result + b) div 2
|
||||
if a[mid] < key: result = mid + 1
|
||||
else: b = mid
|
||||
if result >= len(a) or a[result] != key: result = -1
|
||||
binarySearch(a, key, cmp[T])
|
||||
|
||||
proc smartBinarySearch*[T](a: openArray[T], key: T): int {.deprecated.} =
|
||||
## **Deprecated since version 0.18.1**; Use ``binarySearch`` instead.
|
||||
binarySearch(a,key)
|
||||
binarySearch(a, key, cmp[T])
|
||||
|
||||
const
|
||||
onlySafeCode = true
|
||||
@@ -108,7 +130,7 @@ proc lowerBound*[T, K](a: openArray[T], key: K, cmp: proc(x: T, k: K): int {.clo
|
||||
var count = a.high - a.low + 1
|
||||
var step, pos: int
|
||||
while count != 0:
|
||||
step = count div 2
|
||||
step = count shr 1
|
||||
pos = result + step
|
||||
if cmp(a[pos], key) < 0:
|
||||
result = pos + 1
|
||||
@@ -520,15 +542,15 @@ when isMainModule:
|
||||
|
||||
block testBinarySearch:
|
||||
var noData: seq[int]
|
||||
doAssert binarySearch(noData, 7) == -1
|
||||
doAssert binarySearch(noData, 7) == -1
|
||||
let oneData = @[1]
|
||||
doAssert binarySearch(oneData, 1) == 0
|
||||
doAssert binarySearch(onedata, 7) == -1
|
||||
doAssert binarySearch(oneData, 1) == 0
|
||||
doAssert binarySearch(onedata, 7) == -1
|
||||
let someData = @[1,3,4,7]
|
||||
doAssert binarySearch(someData, 1) == 0
|
||||
doAssert binarySearch(somedata, 7) == 3
|
||||
doAssert binarySearch(someData, 1) == 0
|
||||
doAssert binarySearch(somedata, 7) == 3
|
||||
doAssert binarySearch(someData, -1) == -1
|
||||
doAssert binarySearch(someData, 5) == -1
|
||||
doAssert binarySearch(someData, 5) == -1
|
||||
doAssert binarySearch(someData, 13) == -1
|
||||
let moreData = @[1,3,5,7,4711]
|
||||
doAssert binarySearch(moreData, -1) == -1
|
||||
|
||||
@@ -163,13 +163,13 @@ proc containsOrIncl*(c: var CritBitTree[void], key: string): bool =
|
||||
var n = rawInsert(c, key)
|
||||
result = c.count == oldCount
|
||||
|
||||
proc inc*(c: var CritBitTree[int]; key: string) =
|
||||
## counts the 'key'.
|
||||
proc inc*(c: var CritBitTree[int]; key: string, val: int = 1) =
|
||||
## increments `c[key]` by `val`.
|
||||
let oldCount = c.count
|
||||
var n = rawInsert(c, key)
|
||||
if c.count == oldCount:
|
||||
if c.count == oldCount or oldCount == 0:
|
||||
# not a new key:
|
||||
inc n.val
|
||||
inc n.val, val
|
||||
|
||||
proc incl*(c: var CritBitTree[void], key: string) =
|
||||
## includes `key` in `c`.
|
||||
@@ -352,3 +352,13 @@ when isMainModule:
|
||||
assert toSeq(r.items) == @["abc", "definition", "prefix", "xyz"]
|
||||
|
||||
assert toSeq(r.itemsWithPrefix("de")) == @["definition"]
|
||||
var c = CritBitTree[int]()
|
||||
|
||||
c.inc("a")
|
||||
assert c["a"] == 1
|
||||
|
||||
c.inc("a", 4)
|
||||
assert c["a"] == 5
|
||||
|
||||
c.inc("a", -5)
|
||||
assert c["a"] == 0
|
||||
|
||||
@@ -89,506 +89,21 @@
|
||||
## echo j2
|
||||
|
||||
import
|
||||
hashes, tables, strutils, lexbase, streams, unicode, macros
|
||||
hashes, tables, strutils, lexbase, streams, unicode, macros, parsejson
|
||||
|
||||
export
|
||||
tables.`$`
|
||||
|
||||
export
|
||||
parsejson.JsonEventKind, parsejson.JsonError, JsonParser, JsonKindError,
|
||||
open, close, str, getInt, getFloat, kind, getColumn, getLine, getFilename,
|
||||
errorMsg, errorMsgExpected, next, JsonParsingError, raiseParseErr
|
||||
|
||||
when defined(nimJsonGet):
|
||||
{.pragma: deprecatedGet, deprecated.}
|
||||
else:
|
||||
{.pragma: deprecatedGet.}
|
||||
|
||||
type
|
||||
JsonEventKind* = enum ## enumeration of all events that may occur when parsing
|
||||
jsonError, ## an error occurred during parsing
|
||||
jsonEof, ## end of file reached
|
||||
jsonString, ## a string literal
|
||||
jsonInt, ## an integer literal
|
||||
jsonFloat, ## a float literal
|
||||
jsonTrue, ## the value ``true``
|
||||
jsonFalse, ## the value ``false``
|
||||
jsonNull, ## the value ``null``
|
||||
jsonObjectStart, ## start of an object: the ``{`` token
|
||||
jsonObjectEnd, ## end of an object: the ``}`` token
|
||||
jsonArrayStart, ## start of an array: the ``[`` token
|
||||
jsonArrayEnd ## start of an array: the ``]`` token
|
||||
|
||||
TokKind = enum # must be synchronized with TJsonEventKind!
|
||||
tkError,
|
||||
tkEof,
|
||||
tkString,
|
||||
tkInt,
|
||||
tkFloat,
|
||||
tkTrue,
|
||||
tkFalse,
|
||||
tkNull,
|
||||
tkCurlyLe,
|
||||
tkCurlyRi,
|
||||
tkBracketLe,
|
||||
tkBracketRi,
|
||||
tkColon,
|
||||
tkComma
|
||||
|
||||
JsonError* = enum ## enumeration that lists all errors that can occur
|
||||
errNone, ## no error
|
||||
errInvalidToken, ## invalid token
|
||||
errStringExpected, ## string expected
|
||||
errColonExpected, ## ``:`` expected
|
||||
errCommaExpected, ## ``,`` expected
|
||||
errBracketRiExpected, ## ``]`` expected
|
||||
errCurlyRiExpected, ## ``}`` expected
|
||||
errQuoteExpected, ## ``"`` or ``'`` expected
|
||||
errEOC_Expected, ## ``*/`` expected
|
||||
errEofExpected, ## EOF expected
|
||||
errExprExpected ## expr expected
|
||||
|
||||
ParserState = enum
|
||||
stateEof, stateStart, stateObject, stateArray, stateExpectArrayComma,
|
||||
stateExpectObjectComma, stateExpectColon, stateExpectValue
|
||||
|
||||
JsonParser* = object of BaseLexer ## the parser object.
|
||||
a: string
|
||||
tok: TokKind
|
||||
kind: JsonEventKind
|
||||
err: JsonError
|
||||
state: seq[ParserState]
|
||||
filename: string
|
||||
|
||||
JsonKindError* = object of ValueError ## raised by the ``to`` macro if the
|
||||
## JSON kind is incorrect.
|
||||
|
||||
const
|
||||
errorMessages: array[JsonError, string] = [
|
||||
"no error",
|
||||
"invalid token",
|
||||
"string expected",
|
||||
"':' expected",
|
||||
"',' expected",
|
||||
"']' expected",
|
||||
"'}' expected",
|
||||
"'\"' or \"'\" expected",
|
||||
"'*/' expected",
|
||||
"EOF expected",
|
||||
"expression expected"
|
||||
]
|
||||
tokToStr: array[TokKind, string] = [
|
||||
"invalid token",
|
||||
"EOF",
|
||||
"string literal",
|
||||
"int literal",
|
||||
"float literal",
|
||||
"true",
|
||||
"false",
|
||||
"null",
|
||||
"{", "}", "[", "]", ":", ","
|
||||
]
|
||||
|
||||
proc open*(my: var JsonParser, input: Stream, filename: string) =
|
||||
## initializes the parser with an input stream. `Filename` is only used
|
||||
## for nice error messages.
|
||||
lexbase.open(my, input)
|
||||
my.filename = filename
|
||||
my.state = @[stateStart]
|
||||
my.kind = jsonError
|
||||
my.a = ""
|
||||
|
||||
proc close*(my: var JsonParser) {.inline.} =
|
||||
## closes the parser `my` and its associated input stream.
|
||||
lexbase.close(my)
|
||||
|
||||
proc str*(my: JsonParser): string {.inline.} =
|
||||
## returns the character data for the events: ``jsonInt``, ``jsonFloat``,
|
||||
## ``jsonString``
|
||||
assert(my.kind in {jsonInt, jsonFloat, jsonString})
|
||||
return my.a
|
||||
|
||||
proc getInt*(my: JsonParser): BiggestInt {.inline.} =
|
||||
## returns the number for the event: ``jsonInt``
|
||||
assert(my.kind == jsonInt)
|
||||
return parseBiggestInt(my.a)
|
||||
|
||||
proc getFloat*(my: JsonParser): float {.inline.} =
|
||||
## returns the number for the event: ``jsonFloat``
|
||||
assert(my.kind == jsonFloat)
|
||||
return parseFloat(my.a)
|
||||
|
||||
proc kind*(my: JsonParser): JsonEventKind {.inline.} =
|
||||
## returns the current event type for the JSON parser
|
||||
return my.kind
|
||||
|
||||
proc getColumn*(my: JsonParser): int {.inline.} =
|
||||
## get the current column the parser has arrived at.
|
||||
result = getColNumber(my, my.bufpos)
|
||||
|
||||
proc getLine*(my: JsonParser): int {.inline.} =
|
||||
## get the current line the parser has arrived at.
|
||||
result = my.lineNumber
|
||||
|
||||
proc getFilename*(my: JsonParser): string {.inline.} =
|
||||
## get the filename of the file that the parser processes.
|
||||
result = my.filename
|
||||
|
||||
proc errorMsg*(my: JsonParser): string =
|
||||
## returns a helpful error message for the event ``jsonError``
|
||||
assert(my.kind == jsonError)
|
||||
result = "$1($2, $3) Error: $4" % [
|
||||
my.filename, $getLine(my), $getColumn(my), errorMessages[my.err]]
|
||||
|
||||
proc errorMsgExpected*(my: JsonParser, e: string): string =
|
||||
## returns an error message "`e` expected" in the same format as the
|
||||
## other error messages
|
||||
result = "$1($2, $3) Error: $4" % [
|
||||
my.filename, $getLine(my), $getColumn(my), e & " expected"]
|
||||
|
||||
proc handleHexChar(c: char, x: var int): bool =
|
||||
result = true # Success
|
||||
case c
|
||||
of '0'..'9': x = (x shl 4) or (ord(c) - ord('0'))
|
||||
of 'a'..'f': x = (x shl 4) or (ord(c) - ord('a') + 10)
|
||||
of 'A'..'F': x = (x shl 4) or (ord(c) - ord('A') + 10)
|
||||
else: result = false # error
|
||||
|
||||
proc parseEscapedUTF16(buf: cstring, pos: var int): int =
|
||||
result = 0
|
||||
#UTF-16 escape is always 4 bytes.
|
||||
for _ in 0..3:
|
||||
if handleHexChar(buf[pos], result):
|
||||
inc(pos)
|
||||
else:
|
||||
return -1
|
||||
|
||||
proc parseString(my: var JsonParser): TokKind =
|
||||
result = tkString
|
||||
var pos = my.bufpos + 1
|
||||
var buf = my.buf
|
||||
while true:
|
||||
case buf[pos]
|
||||
of '\0':
|
||||
my.err = errQuoteExpected
|
||||
result = tkError
|
||||
break
|
||||
of '"':
|
||||
inc(pos)
|
||||
break
|
||||
of '\\':
|
||||
case buf[pos+1]
|
||||
of '\\', '"', '\'', '/':
|
||||
add(my.a, buf[pos+1])
|
||||
inc(pos, 2)
|
||||
of 'b':
|
||||
add(my.a, '\b')
|
||||
inc(pos, 2)
|
||||
of 'f':
|
||||
add(my.a, '\f')
|
||||
inc(pos, 2)
|
||||
of 'n':
|
||||
add(my.a, '\L')
|
||||
inc(pos, 2)
|
||||
of 'r':
|
||||
add(my.a, '\C')
|
||||
inc(pos, 2)
|
||||
of 't':
|
||||
add(my.a, '\t')
|
||||
inc(pos, 2)
|
||||
of 'u':
|
||||
inc(pos, 2)
|
||||
var r = parseEscapedUTF16(buf, pos)
|
||||
if r < 0:
|
||||
my.err = errInvalidToken
|
||||
break
|
||||
# Deal with surrogates
|
||||
if (r and 0xfc00) == 0xd800:
|
||||
if buf[pos] & buf[pos+1] != "\\u":
|
||||
my.err = errInvalidToken
|
||||
break
|
||||
inc(pos, 2)
|
||||
var s = parseEscapedUTF16(buf, pos)
|
||||
if (s and 0xfc00) == 0xdc00 and s > 0:
|
||||
r = 0x10000 + (((r - 0xd800) shl 10) or (s - 0xdc00))
|
||||
else:
|
||||
my.err = errInvalidToken
|
||||
break
|
||||
add(my.a, toUTF8(Rune(r)))
|
||||
else:
|
||||
# don't bother with the error
|
||||
add(my.a, buf[pos])
|
||||
inc(pos)
|
||||
of '\c':
|
||||
pos = lexbase.handleCR(my, pos)
|
||||
buf = my.buf
|
||||
add(my.a, '\c')
|
||||
of '\L':
|
||||
pos = lexbase.handleLF(my, pos)
|
||||
buf = my.buf
|
||||
add(my.a, '\L')
|
||||
else:
|
||||
add(my.a, buf[pos])
|
||||
inc(pos)
|
||||
my.bufpos = pos # store back
|
||||
|
||||
proc skip(my: var JsonParser) =
|
||||
var pos = my.bufpos
|
||||
var buf = my.buf
|
||||
while true:
|
||||
case buf[pos]
|
||||
of '/':
|
||||
if buf[pos+1] == '/':
|
||||
# skip line comment:
|
||||
inc(pos, 2)
|
||||
while true:
|
||||
case buf[pos]
|
||||
of '\0':
|
||||
break
|
||||
of '\c':
|
||||
pos = lexbase.handleCR(my, pos)
|
||||
buf = my.buf
|
||||
break
|
||||
of '\L':
|
||||
pos = lexbase.handleLF(my, pos)
|
||||
buf = my.buf
|
||||
break
|
||||
else:
|
||||
inc(pos)
|
||||
elif buf[pos+1] == '*':
|
||||
# skip long comment:
|
||||
inc(pos, 2)
|
||||
while true:
|
||||
case buf[pos]
|
||||
of '\0':
|
||||
my.err = errEOC_Expected
|
||||
break
|
||||
of '\c':
|
||||
pos = lexbase.handleCR(my, pos)
|
||||
buf = my.buf
|
||||
of '\L':
|
||||
pos = lexbase.handleLF(my, pos)
|
||||
buf = my.buf
|
||||
of '*':
|
||||
inc(pos)
|
||||
if buf[pos] == '/':
|
||||
inc(pos)
|
||||
break
|
||||
else:
|
||||
inc(pos)
|
||||
else:
|
||||
break
|
||||
of ' ', '\t':
|
||||
inc(pos)
|
||||
of '\c':
|
||||
pos = lexbase.handleCR(my, pos)
|
||||
buf = my.buf
|
||||
of '\L':
|
||||
pos = lexbase.handleLF(my, pos)
|
||||
buf = my.buf
|
||||
else:
|
||||
break
|
||||
my.bufpos = pos
|
||||
|
||||
proc parseNumber(my: var JsonParser) =
|
||||
var pos = my.bufpos
|
||||
var buf = my.buf
|
||||
if buf[pos] == '-':
|
||||
add(my.a, '-')
|
||||
inc(pos)
|
||||
if buf[pos] == '.':
|
||||
add(my.a, "0.")
|
||||
inc(pos)
|
||||
else:
|
||||
while buf[pos] in Digits:
|
||||
add(my.a, buf[pos])
|
||||
inc(pos)
|
||||
if buf[pos] == '.':
|
||||
add(my.a, '.')
|
||||
inc(pos)
|
||||
# digits after the dot:
|
||||
while buf[pos] in Digits:
|
||||
add(my.a, buf[pos])
|
||||
inc(pos)
|
||||
if buf[pos] in {'E', 'e'}:
|
||||
add(my.a, buf[pos])
|
||||
inc(pos)
|
||||
if buf[pos] in {'+', '-'}:
|
||||
add(my.a, buf[pos])
|
||||
inc(pos)
|
||||
while buf[pos] in Digits:
|
||||
add(my.a, buf[pos])
|
||||
inc(pos)
|
||||
my.bufpos = pos
|
||||
|
||||
proc parseName(my: var JsonParser) =
|
||||
var pos = my.bufpos
|
||||
var buf = my.buf
|
||||
if buf[pos] in IdentStartChars:
|
||||
while buf[pos] in IdentChars:
|
||||
add(my.a, buf[pos])
|
||||
inc(pos)
|
||||
my.bufpos = pos
|
||||
|
||||
proc getTok(my: var JsonParser): TokKind =
|
||||
setLen(my.a, 0)
|
||||
skip(my) # skip whitespace, comments
|
||||
case my.buf[my.bufpos]
|
||||
of '-', '.', '0'..'9':
|
||||
parseNumber(my)
|
||||
if {'.', 'e', 'E'} in my.a:
|
||||
result = tkFloat
|
||||
else:
|
||||
result = tkInt
|
||||
of '"':
|
||||
result = parseString(my)
|
||||
of '[':
|
||||
inc(my.bufpos)
|
||||
result = tkBracketLe
|
||||
of '{':
|
||||
inc(my.bufpos)
|
||||
result = tkCurlyLe
|
||||
of ']':
|
||||
inc(my.bufpos)
|
||||
result = tkBracketRi
|
||||
of '}':
|
||||
inc(my.bufpos)
|
||||
result = tkCurlyRi
|
||||
of ',':
|
||||
inc(my.bufpos)
|
||||
result = tkComma
|
||||
of ':':
|
||||
inc(my.bufpos)
|
||||
result = tkColon
|
||||
of '\0':
|
||||
result = tkEof
|
||||
of 'a'..'z', 'A'..'Z', '_':
|
||||
parseName(my)
|
||||
case my.a
|
||||
of "null": result = tkNull
|
||||
of "true": result = tkTrue
|
||||
of "false": result = tkFalse
|
||||
else: result = tkError
|
||||
else:
|
||||
inc(my.bufpos)
|
||||
result = tkError
|
||||
my.tok = result
|
||||
|
||||
proc next*(my: var JsonParser) =
|
||||
## retrieves the first/next event. This controls the parser.
|
||||
var tk = getTok(my)
|
||||
var i = my.state.len-1
|
||||
# the following code is a state machine. If we had proper coroutines,
|
||||
# the code could be much simpler.
|
||||
case my.state[i]
|
||||
of stateEof:
|
||||
if tk == tkEof:
|
||||
my.kind = jsonEof
|
||||
else:
|
||||
my.kind = jsonError
|
||||
my.err = errEofExpected
|
||||
of stateStart:
|
||||
# tokens allowed?
|
||||
case tk
|
||||
of tkString, tkInt, tkFloat, tkTrue, tkFalse, tkNull:
|
||||
my.state[i] = stateEof # expect EOF next!
|
||||
my.kind = JsonEventKind(ord(tk))
|
||||
of tkBracketLe:
|
||||
my.state.add(stateArray) # we expect any
|
||||
my.kind = jsonArrayStart
|
||||
of tkCurlyLe:
|
||||
my.state.add(stateObject)
|
||||
my.kind = jsonObjectStart
|
||||
of tkEof:
|
||||
my.kind = jsonEof
|
||||
else:
|
||||
my.kind = jsonError
|
||||
my.err = errEofExpected
|
||||
of stateObject:
|
||||
case tk
|
||||
of tkString, tkInt, tkFloat, tkTrue, tkFalse, tkNull:
|
||||
my.state.add(stateExpectColon)
|
||||
my.kind = JsonEventKind(ord(tk))
|
||||
of tkBracketLe:
|
||||
my.state.add(stateExpectColon)
|
||||
my.state.add(stateArray)
|
||||
my.kind = jsonArrayStart
|
||||
of tkCurlyLe:
|
||||
my.state.add(stateExpectColon)
|
||||
my.state.add(stateObject)
|
||||
my.kind = jsonObjectStart
|
||||
of tkCurlyRi:
|
||||
my.kind = jsonObjectEnd
|
||||
discard my.state.pop()
|
||||
else:
|
||||
my.kind = jsonError
|
||||
my.err = errCurlyRiExpected
|
||||
of stateArray:
|
||||
case tk
|
||||
of tkString, tkInt, tkFloat, tkTrue, tkFalse, tkNull:
|
||||
my.state.add(stateExpectArrayComma) # expect value next!
|
||||
my.kind = JsonEventKind(ord(tk))
|
||||
of tkBracketLe:
|
||||
my.state.add(stateExpectArrayComma)
|
||||
my.state.add(stateArray)
|
||||
my.kind = jsonArrayStart
|
||||
of tkCurlyLe:
|
||||
my.state.add(stateExpectArrayComma)
|
||||
my.state.add(stateObject)
|
||||
my.kind = jsonObjectStart
|
||||
of tkBracketRi:
|
||||
my.kind = jsonArrayEnd
|
||||
discard my.state.pop()
|
||||
else:
|
||||
my.kind = jsonError
|
||||
my.err = errBracketRiExpected
|
||||
of stateExpectArrayComma:
|
||||
case tk
|
||||
of tkComma:
|
||||
discard my.state.pop()
|
||||
next(my)
|
||||
of tkBracketRi:
|
||||
my.kind = jsonArrayEnd
|
||||
discard my.state.pop() # pop stateExpectArrayComma
|
||||
discard my.state.pop() # pop stateArray
|
||||
else:
|
||||
my.kind = jsonError
|
||||
my.err = errBracketRiExpected
|
||||
of stateExpectObjectComma:
|
||||
case tk
|
||||
of tkComma:
|
||||
discard my.state.pop()
|
||||
next(my)
|
||||
of tkCurlyRi:
|
||||
my.kind = jsonObjectEnd
|
||||
discard my.state.pop() # pop stateExpectObjectComma
|
||||
discard my.state.pop() # pop stateObject
|
||||
else:
|
||||
my.kind = jsonError
|
||||
my.err = errCurlyRiExpected
|
||||
of stateExpectColon:
|
||||
case tk
|
||||
of tkColon:
|
||||
my.state[i] = stateExpectValue
|
||||
next(my)
|
||||
else:
|
||||
my.kind = jsonError
|
||||
my.err = errColonExpected
|
||||
of stateExpectValue:
|
||||
case tk
|
||||
of tkString, tkInt, tkFloat, tkTrue, tkFalse, tkNull:
|
||||
my.state[i] = stateExpectObjectComma
|
||||
my.kind = JsonEventKind(ord(tk))
|
||||
of tkBracketLe:
|
||||
my.state[i] = stateExpectObjectComma
|
||||
my.state.add(stateArray)
|
||||
my.kind = jsonArrayStart
|
||||
of tkCurlyLe:
|
||||
my.state[i] = stateExpectObjectComma
|
||||
my.state.add(stateObject)
|
||||
my.kind = jsonObjectStart
|
||||
else:
|
||||
my.kind = jsonError
|
||||
my.err = errExprExpected
|
||||
|
||||
|
||||
# ------------- higher level interface ---------------------------------------
|
||||
|
||||
type
|
||||
JsonNodeKind* = enum ## possible JSON node types
|
||||
JNull,
|
||||
@@ -617,12 +132,6 @@ type
|
||||
of JArray:
|
||||
elems*: seq[JsonNode]
|
||||
|
||||
JsonParsingError* = object of ValueError ## is raised for a JSON error
|
||||
|
||||
proc raiseParseErr*(p: JsonParser, msg: string) {.noinline, noreturn.} =
|
||||
## raises an `EJsonParsingError` exception.
|
||||
raise newException(JsonParsingError, errorMsgExpected(p, msg))
|
||||
|
||||
proc newJString*(s: string): JsonNode =
|
||||
## Creates a new `JString JsonNode`.
|
||||
new(result)
|
||||
@@ -1000,7 +509,7 @@ proc delete*(obj: JsonNode, key: string) =
|
||||
## Deletes ``obj[key]``.
|
||||
assert(obj.kind == JObject)
|
||||
if not obj.fields.hasKey(key):
|
||||
raise newException(IndexError, "key not in object")
|
||||
raise newException(KeyError, "key not in object")
|
||||
obj.fields.del(key)
|
||||
|
||||
proc copy*(p: JsonNode): JsonNode =
|
||||
@@ -1194,10 +703,6 @@ iterator mpairs*(node: var JsonNode): tuple[key: string, val: var JsonNode] =
|
||||
for key, val in mpairs(node.fields):
|
||||
yield (key, val)
|
||||
|
||||
proc eat(p: var JsonParser, tok: TokKind) =
|
||||
if p.tok == tok: discard getTok(p)
|
||||
else: raiseParseErr(p, tokToStr[tok])
|
||||
|
||||
proc parseJson(p: var JsonParser): JsonNode =
|
||||
## Parses JSON from a JSON Parser `p`.
|
||||
case p.tok
|
||||
@@ -1253,10 +758,12 @@ when not defined(js):
|
||||
## If `s` contains extra data, it will raise `JsonParsingError`.
|
||||
var p: JsonParser
|
||||
p.open(s, filename)
|
||||
defer: p.close()
|
||||
discard getTok(p) # read first token
|
||||
result = p.parseJson()
|
||||
eat(p, tkEof) # check if there is no extra data
|
||||
try:
|
||||
discard getTok(p) # read first token
|
||||
result = p.parseJson()
|
||||
eat(p, tkEof) # check if there is no extra data
|
||||
finally:
|
||||
p.close()
|
||||
|
||||
proc parseJson*(buffer: string): JsonNode =
|
||||
## Parses JSON from `buffer`.
|
||||
@@ -1989,18 +1496,18 @@ when isMainModule:
|
||||
# Bounds checking
|
||||
try:
|
||||
let a = testJson["a"][9]
|
||||
doAssert(false, "EInvalidIndex not thrown")
|
||||
doAssert(false, "IndexError not thrown")
|
||||
except IndexError:
|
||||
discard
|
||||
try:
|
||||
let a = testJson["a"][-1]
|
||||
doAssert(false, "EInvalidIndex not thrown")
|
||||
doAssert(false, "IndexError not thrown")
|
||||
except IndexError:
|
||||
discard
|
||||
try:
|
||||
doAssert(testJson["a"][0].num == 1, "Index doesn't correspond to its value")
|
||||
except:
|
||||
doAssert(false, "EInvalidIndex thrown for valid index")
|
||||
doAssert(false, "IndexError thrown for valid index")
|
||||
|
||||
doAssert(testJson{"b"}.str=="asd", "Couldn't fetch a singly nested key with {}")
|
||||
doAssert(isNil(testJson{"nonexistent"}), "Non-existent keys should return nil")
|
||||
|
||||
@@ -41,7 +41,7 @@ proc fac*(n: int): int =
|
||||
createFactTable[13]()
|
||||
else:
|
||||
createFactTable[21]()
|
||||
assert(n > 0, $n & " must not be negative.")
|
||||
assert(n >= 0, $n & " must not be negative.")
|
||||
assert(n < factTable.len, $n & " is too large to look up in the table")
|
||||
factTable[n]
|
||||
|
||||
@@ -560,3 +560,14 @@ when isMainModule:
|
||||
assert sgn(Inf) == 1
|
||||
assert sgn(NaN) == 0
|
||||
|
||||
block: # fac() tests
|
||||
try:
|
||||
discard fac(-1)
|
||||
except AssertionError:
|
||||
discard
|
||||
|
||||
doAssert fac(0) == 1
|
||||
doAssert fac(1) == 1
|
||||
doAssert fac(2) == 2
|
||||
doAssert fac(3) == 6
|
||||
doAssert fac(4) == 24
|
||||
|
||||
@@ -139,6 +139,7 @@ proc findExe*(exe: string, followSymlinks: bool = true;
|
||||
## is added the `ExeExts <#ExeExts>`_ file extensions if it has none.
|
||||
## If the system supports symlinks it also resolves them until it
|
||||
## meets the actual file. This behavior can be disabled if desired.
|
||||
if exe.len == 0: return
|
||||
template checkCurrentDir() =
|
||||
for ext in extensions:
|
||||
result = addFileExt(exe, ext)
|
||||
@@ -149,6 +150,7 @@ proc findExe*(exe: string, followSymlinks: bool = true;
|
||||
checkCurrentDir()
|
||||
let path = string(getEnv("PATH"))
|
||||
for candidate in split(path, PathSep):
|
||||
if candidate.len == 0: continue
|
||||
when defined(windows):
|
||||
var x = (if candidate[0] == '"' and candidate[^1] == '"':
|
||||
substr(candidate, 1, candidate.len-2) else: candidate) /
|
||||
@@ -1649,4 +1651,4 @@ proc setLastModificationTime*(file: string, t: times.Time) =
|
||||
var ft = t.toWinTime.toFILETIME
|
||||
let res = setFileTime(h, nil, nil, ft.addr)
|
||||
discard h.closeHandle
|
||||
if res == 0'i32: raiseOSError(osLastError())
|
||||
if res == 0'i32: raiseOSError(osLastError())
|
||||
|
||||
535
lib/pure/parsejson.nim
Normal file
535
lib/pure/parsejson.nim
Normal file
@@ -0,0 +1,535 @@
|
||||
#
|
||||
#
|
||||
# Nim's Runtime Library
|
||||
# (c) Copyright 2018 Nim contributors
|
||||
#
|
||||
# See the file "copying.txt", included in this
|
||||
# distribution, for details about the copyright.
|
||||
#
|
||||
|
||||
## This module implements a json parser. It is used
|
||||
## and exported by the ``json`` standard library
|
||||
## module, but can also be used in its own right.
|
||||
|
||||
import
|
||||
strutils, lexbase, streams, unicode
|
||||
|
||||
type
|
||||
JsonEventKind* = enum ## enumeration of all events that may occur when parsing
|
||||
jsonError, ## an error occurred during parsing
|
||||
jsonEof, ## end of file reached
|
||||
jsonString, ## a string literal
|
||||
jsonInt, ## an integer literal
|
||||
jsonFloat, ## a float literal
|
||||
jsonTrue, ## the value ``true``
|
||||
jsonFalse, ## the value ``false``
|
||||
jsonNull, ## the value ``null``
|
||||
jsonObjectStart, ## start of an object: the ``{`` token
|
||||
jsonObjectEnd, ## end of an object: the ``}`` token
|
||||
jsonArrayStart, ## start of an array: the ``[`` token
|
||||
jsonArrayEnd ## start of an array: the ``]`` token
|
||||
|
||||
TokKind* = enum # must be synchronized with TJsonEventKind!
|
||||
tkError,
|
||||
tkEof,
|
||||
tkString,
|
||||
tkInt,
|
||||
tkFloat,
|
||||
tkTrue,
|
||||
tkFalse,
|
||||
tkNull,
|
||||
tkCurlyLe,
|
||||
tkCurlyRi,
|
||||
tkBracketLe,
|
||||
tkBracketRi,
|
||||
tkColon,
|
||||
tkComma
|
||||
|
||||
JsonError* = enum ## enumeration that lists all errors that can occur
|
||||
errNone, ## no error
|
||||
errInvalidToken, ## invalid token
|
||||
errStringExpected, ## string expected
|
||||
errColonExpected, ## ``:`` expected
|
||||
errCommaExpected, ## ``,`` expected
|
||||
errBracketRiExpected, ## ``]`` expected
|
||||
errCurlyRiExpected, ## ``}`` expected
|
||||
errQuoteExpected, ## ``"`` or ``'`` expected
|
||||
errEOC_Expected, ## ``*/`` expected
|
||||
errEofExpected, ## EOF expected
|
||||
errExprExpected ## expr expected
|
||||
|
||||
ParserState = enum
|
||||
stateEof, stateStart, stateObject, stateArray, stateExpectArrayComma,
|
||||
stateExpectObjectComma, stateExpectColon, stateExpectValue
|
||||
|
||||
JsonParser* = object of BaseLexer ## the parser object.
|
||||
a*: string
|
||||
tok*: TokKind
|
||||
kind: JsonEventKind
|
||||
err: JsonError
|
||||
state: seq[ParserState]
|
||||
filename: string
|
||||
rawStringLiterals: bool
|
||||
|
||||
JsonKindError* = object of ValueError ## raised by the ``to`` macro if the
|
||||
## JSON kind is incorrect.
|
||||
JsonParsingError* = object of ValueError ## is raised for a JSON error
|
||||
|
||||
const
|
||||
errorMessages*: array[JsonError, string] = [
|
||||
"no error",
|
||||
"invalid token",
|
||||
"string expected",
|
||||
"':' expected",
|
||||
"',' expected",
|
||||
"']' expected",
|
||||
"'}' expected",
|
||||
"'\"' or \"'\" expected",
|
||||
"'*/' expected",
|
||||
"EOF expected",
|
||||
"expression expected"
|
||||
]
|
||||
tokToStr: array[TokKind, string] = [
|
||||
"invalid token",
|
||||
"EOF",
|
||||
"string literal",
|
||||
"int literal",
|
||||
"float literal",
|
||||
"true",
|
||||
"false",
|
||||
"null",
|
||||
"{", "}", "[", "]", ":", ","
|
||||
]
|
||||
|
||||
proc open*(my: var JsonParser, input: Stream, filename: string;
|
||||
rawStringLiterals = false) =
|
||||
## initializes the parser with an input stream. `Filename` is only used
|
||||
## for nice error messages. If `rawStringLiterals` is true, string literals
|
||||
## are kepts with their surrounding quotes and escape sequences in them are
|
||||
## left untouched too.
|
||||
lexbase.open(my, input)
|
||||
my.filename = filename
|
||||
my.state = @[stateStart]
|
||||
my.kind = jsonError
|
||||
my.a = ""
|
||||
my.rawStringLiterals = rawStringLiterals
|
||||
|
||||
proc close*(my: var JsonParser) {.inline.} =
|
||||
## closes the parser `my` and its associated input stream.
|
||||
lexbase.close(my)
|
||||
|
||||
proc str*(my: JsonParser): string {.inline.} =
|
||||
## returns the character data for the events: ``jsonInt``, ``jsonFloat``,
|
||||
## ``jsonString``
|
||||
assert(my.kind in {jsonInt, jsonFloat, jsonString})
|
||||
return my.a
|
||||
|
||||
proc getInt*(my: JsonParser): BiggestInt {.inline.} =
|
||||
## returns the number for the event: ``jsonInt``
|
||||
assert(my.kind == jsonInt)
|
||||
return parseBiggestInt(my.a)
|
||||
|
||||
proc getFloat*(my: JsonParser): float {.inline.} =
|
||||
## returns the number for the event: ``jsonFloat``
|
||||
assert(my.kind == jsonFloat)
|
||||
return parseFloat(my.a)
|
||||
|
||||
proc kind*(my: JsonParser): JsonEventKind {.inline.} =
|
||||
## returns the current event type for the JSON parser
|
||||
return my.kind
|
||||
|
||||
proc getColumn*(my: JsonParser): int {.inline.} =
|
||||
## get the current column the parser has arrived at.
|
||||
result = getColNumber(my, my.bufpos)
|
||||
|
||||
proc getLine*(my: JsonParser): int {.inline.} =
|
||||
## get the current line the parser has arrived at.
|
||||
result = my.lineNumber
|
||||
|
||||
proc getFilename*(my: JsonParser): string {.inline.} =
|
||||
## get the filename of the file that the parser processes.
|
||||
result = my.filename
|
||||
|
||||
proc errorMsg*(my: JsonParser): string =
|
||||
## returns a helpful error message for the event ``jsonError``
|
||||
assert(my.kind == jsonError)
|
||||
result = "$1($2, $3) Error: $4" % [
|
||||
my.filename, $getLine(my), $getColumn(my), errorMessages[my.err]]
|
||||
|
||||
proc errorMsgExpected*(my: JsonParser, e: string): string =
|
||||
## returns an error message "`e` expected" in the same format as the
|
||||
## other error messages
|
||||
result = "$1($2, $3) Error: $4" % [
|
||||
my.filename, $getLine(my), $getColumn(my), e & " expected"]
|
||||
|
||||
proc handleHexChar(c: char, x: var int): bool =
|
||||
result = true # Success
|
||||
case c
|
||||
of '0'..'9': x = (x shl 4) or (ord(c) - ord('0'))
|
||||
of 'a'..'f': x = (x shl 4) or (ord(c) - ord('a') + 10)
|
||||
of 'A'..'F': x = (x shl 4) or (ord(c) - ord('A') + 10)
|
||||
else: result = false # error
|
||||
|
||||
proc parseEscapedUTF16*(buf: cstring, pos: var int): int =
|
||||
result = 0
|
||||
#UTF-16 escape is always 4 bytes.
|
||||
for _ in 0..3:
|
||||
if handleHexChar(buf[pos], result):
|
||||
inc(pos)
|
||||
else:
|
||||
return -1
|
||||
|
||||
proc parseString(my: var JsonParser): TokKind =
|
||||
result = tkString
|
||||
var pos = my.bufpos + 1
|
||||
var buf = my.buf
|
||||
if my.rawStringLiterals:
|
||||
add(my.a, '"')
|
||||
while true:
|
||||
case buf[pos]
|
||||
of '\0':
|
||||
my.err = errQuoteExpected
|
||||
result = tkError
|
||||
break
|
||||
of '"':
|
||||
if my.rawStringLiterals:
|
||||
add(my.a, '"')
|
||||
inc(pos)
|
||||
break
|
||||
of '\\':
|
||||
if my.rawStringLiterals:
|
||||
add(my.a, '\\')
|
||||
case buf[pos+1]
|
||||
of '\\', '"', '\'', '/':
|
||||
add(my.a, buf[pos+1])
|
||||
inc(pos, 2)
|
||||
of 'b':
|
||||
add(my.a, '\b')
|
||||
inc(pos, 2)
|
||||
of 'f':
|
||||
add(my.a, '\f')
|
||||
inc(pos, 2)
|
||||
of 'n':
|
||||
add(my.a, '\L')
|
||||
inc(pos, 2)
|
||||
of 'r':
|
||||
add(my.a, '\C')
|
||||
inc(pos, 2)
|
||||
of 't':
|
||||
add(my.a, '\t')
|
||||
inc(pos, 2)
|
||||
of 'u':
|
||||
if my.rawStringLiterals:
|
||||
add(my.a, 'u')
|
||||
inc(pos, 2)
|
||||
var pos2 = pos
|
||||
var r = parseEscapedUTF16(buf, pos)
|
||||
if r < 0:
|
||||
my.err = errInvalidToken
|
||||
break
|
||||
# Deal with surrogates
|
||||
if (r and 0xfc00) == 0xd800:
|
||||
if buf[pos] != '\\' or buf[pos+1] != 'u':
|
||||
my.err = errInvalidToken
|
||||
break
|
||||
inc(pos, 2)
|
||||
var s = parseEscapedUTF16(buf, pos)
|
||||
if (s and 0xfc00) == 0xdc00 and s > 0:
|
||||
r = 0x10000 + (((r - 0xd800) shl 10) or (s - 0xdc00))
|
||||
else:
|
||||
my.err = errInvalidToken
|
||||
break
|
||||
if my.rawStringLiterals:
|
||||
let length = pos - pos2
|
||||
for i in 1 .. length:
|
||||
if buf[pos2] in {'0'..'9', 'A'..'F', 'a'..'f'}:
|
||||
add(my.a, buf[pos2])
|
||||
inc pos2
|
||||
else:
|
||||
break
|
||||
else:
|
||||
add(my.a, toUTF8(Rune(r)))
|
||||
else:
|
||||
# don't bother with the error
|
||||
add(my.a, buf[pos])
|
||||
inc(pos)
|
||||
of '\c':
|
||||
pos = lexbase.handleCR(my, pos)
|
||||
buf = my.buf
|
||||
add(my.a, '\c')
|
||||
of '\L':
|
||||
pos = lexbase.handleLF(my, pos)
|
||||
buf = my.buf
|
||||
add(my.a, '\L')
|
||||
else:
|
||||
add(my.a, buf[pos])
|
||||
inc(pos)
|
||||
my.bufpos = pos # store back
|
||||
|
||||
proc skip(my: var JsonParser) =
|
||||
var pos = my.bufpos
|
||||
var buf = my.buf
|
||||
while true:
|
||||
case buf[pos]
|
||||
of '/':
|
||||
if buf[pos+1] == '/':
|
||||
# skip line comment:
|
||||
inc(pos, 2)
|
||||
while true:
|
||||
case buf[pos]
|
||||
of '\0':
|
||||
break
|
||||
of '\c':
|
||||
pos = lexbase.handleCR(my, pos)
|
||||
buf = my.buf
|
||||
break
|
||||
of '\L':
|
||||
pos = lexbase.handleLF(my, pos)
|
||||
buf = my.buf
|
||||
break
|
||||
else:
|
||||
inc(pos)
|
||||
elif buf[pos+1] == '*':
|
||||
# skip long comment:
|
||||
inc(pos, 2)
|
||||
while true:
|
||||
case buf[pos]
|
||||
of '\0':
|
||||
my.err = errEOC_Expected
|
||||
break
|
||||
of '\c':
|
||||
pos = lexbase.handleCR(my, pos)
|
||||
buf = my.buf
|
||||
of '\L':
|
||||
pos = lexbase.handleLF(my, pos)
|
||||
buf = my.buf
|
||||
of '*':
|
||||
inc(pos)
|
||||
if buf[pos] == '/':
|
||||
inc(pos)
|
||||
break
|
||||
else:
|
||||
inc(pos)
|
||||
else:
|
||||
break
|
||||
of ' ', '\t':
|
||||
inc(pos)
|
||||
of '\c':
|
||||
pos = lexbase.handleCR(my, pos)
|
||||
buf = my.buf
|
||||
of '\L':
|
||||
pos = lexbase.handleLF(my, pos)
|
||||
buf = my.buf
|
||||
else:
|
||||
break
|
||||
my.bufpos = pos
|
||||
|
||||
proc parseNumber(my: var JsonParser) =
|
||||
var pos = my.bufpos
|
||||
var buf = my.buf
|
||||
if buf[pos] == '-':
|
||||
add(my.a, '-')
|
||||
inc(pos)
|
||||
if buf[pos] == '.':
|
||||
add(my.a, "0.")
|
||||
inc(pos)
|
||||
else:
|
||||
while buf[pos] in Digits:
|
||||
add(my.a, buf[pos])
|
||||
inc(pos)
|
||||
if buf[pos] == '.':
|
||||
add(my.a, '.')
|
||||
inc(pos)
|
||||
# digits after the dot:
|
||||
while buf[pos] in Digits:
|
||||
add(my.a, buf[pos])
|
||||
inc(pos)
|
||||
if buf[pos] in {'E', 'e'}:
|
||||
add(my.a, buf[pos])
|
||||
inc(pos)
|
||||
if buf[pos] in {'+', '-'}:
|
||||
add(my.a, buf[pos])
|
||||
inc(pos)
|
||||
while buf[pos] in Digits:
|
||||
add(my.a, buf[pos])
|
||||
inc(pos)
|
||||
my.bufpos = pos
|
||||
|
||||
proc parseName(my: var JsonParser) =
|
||||
var pos = my.bufpos
|
||||
var buf = my.buf
|
||||
if buf[pos] in IdentStartChars:
|
||||
while buf[pos] in IdentChars:
|
||||
add(my.a, buf[pos])
|
||||
inc(pos)
|
||||
my.bufpos = pos
|
||||
|
||||
proc getTok*(my: var JsonParser): TokKind =
|
||||
setLen(my.a, 0)
|
||||
skip(my) # skip whitespace, comments
|
||||
case my.buf[my.bufpos]
|
||||
of '-', '.', '0'..'9':
|
||||
parseNumber(my)
|
||||
if {'.', 'e', 'E'} in my.a:
|
||||
result = tkFloat
|
||||
else:
|
||||
result = tkInt
|
||||
of '"':
|
||||
result = parseString(my)
|
||||
of '[':
|
||||
inc(my.bufpos)
|
||||
result = tkBracketLe
|
||||
of '{':
|
||||
inc(my.bufpos)
|
||||
result = tkCurlyLe
|
||||
of ']':
|
||||
inc(my.bufpos)
|
||||
result = tkBracketRi
|
||||
of '}':
|
||||
inc(my.bufpos)
|
||||
result = tkCurlyRi
|
||||
of ',':
|
||||
inc(my.bufpos)
|
||||
result = tkComma
|
||||
of ':':
|
||||
inc(my.bufpos)
|
||||
result = tkColon
|
||||
of '\0':
|
||||
result = tkEof
|
||||
of 'a'..'z', 'A'..'Z', '_':
|
||||
parseName(my)
|
||||
case my.a
|
||||
of "null": result = tkNull
|
||||
of "true": result = tkTrue
|
||||
of "false": result = tkFalse
|
||||
else: result = tkError
|
||||
else:
|
||||
inc(my.bufpos)
|
||||
result = tkError
|
||||
my.tok = result
|
||||
|
||||
|
||||
proc next*(my: var JsonParser) =
|
||||
## retrieves the first/next event. This controls the parser.
|
||||
var tk = getTok(my)
|
||||
var i = my.state.len-1
|
||||
# the following code is a state machine. If we had proper coroutines,
|
||||
# the code could be much simpler.
|
||||
case my.state[i]
|
||||
of stateEof:
|
||||
if tk == tkEof:
|
||||
my.kind = jsonEof
|
||||
else:
|
||||
my.kind = jsonError
|
||||
my.err = errEofExpected
|
||||
of stateStart:
|
||||
# tokens allowed?
|
||||
case tk
|
||||
of tkString, tkInt, tkFloat, tkTrue, tkFalse, tkNull:
|
||||
my.state[i] = stateEof # expect EOF next!
|
||||
my.kind = JsonEventKind(ord(tk))
|
||||
of tkBracketLe:
|
||||
my.state.add(stateArray) # we expect any
|
||||
my.kind = jsonArrayStart
|
||||
of tkCurlyLe:
|
||||
my.state.add(stateObject)
|
||||
my.kind = jsonObjectStart
|
||||
of tkEof:
|
||||
my.kind = jsonEof
|
||||
else:
|
||||
my.kind = jsonError
|
||||
my.err = errEofExpected
|
||||
of stateObject:
|
||||
case tk
|
||||
of tkString, tkInt, tkFloat, tkTrue, tkFalse, tkNull:
|
||||
my.state.add(stateExpectColon)
|
||||
my.kind = JsonEventKind(ord(tk))
|
||||
of tkBracketLe:
|
||||
my.state.add(stateExpectColon)
|
||||
my.state.add(stateArray)
|
||||
my.kind = jsonArrayStart
|
||||
of tkCurlyLe:
|
||||
my.state.add(stateExpectColon)
|
||||
my.state.add(stateObject)
|
||||
my.kind = jsonObjectStart
|
||||
of tkCurlyRi:
|
||||
my.kind = jsonObjectEnd
|
||||
discard my.state.pop()
|
||||
else:
|
||||
my.kind = jsonError
|
||||
my.err = errCurlyRiExpected
|
||||
of stateArray:
|
||||
case tk
|
||||
of tkString, tkInt, tkFloat, tkTrue, tkFalse, tkNull:
|
||||
my.state.add(stateExpectArrayComma) # expect value next!
|
||||
my.kind = JsonEventKind(ord(tk))
|
||||
of tkBracketLe:
|
||||
my.state.add(stateExpectArrayComma)
|
||||
my.state.add(stateArray)
|
||||
my.kind = jsonArrayStart
|
||||
of tkCurlyLe:
|
||||
my.state.add(stateExpectArrayComma)
|
||||
my.state.add(stateObject)
|
||||
my.kind = jsonObjectStart
|
||||
of tkBracketRi:
|
||||
my.kind = jsonArrayEnd
|
||||
discard my.state.pop()
|
||||
else:
|
||||
my.kind = jsonError
|
||||
my.err = errBracketRiExpected
|
||||
of stateExpectArrayComma:
|
||||
case tk
|
||||
of tkComma:
|
||||
discard my.state.pop()
|
||||
next(my)
|
||||
of tkBracketRi:
|
||||
my.kind = jsonArrayEnd
|
||||
discard my.state.pop() # pop stateExpectArrayComma
|
||||
discard my.state.pop() # pop stateArray
|
||||
else:
|
||||
my.kind = jsonError
|
||||
my.err = errBracketRiExpected
|
||||
of stateExpectObjectComma:
|
||||
case tk
|
||||
of tkComma:
|
||||
discard my.state.pop()
|
||||
next(my)
|
||||
of tkCurlyRi:
|
||||
my.kind = jsonObjectEnd
|
||||
discard my.state.pop() # pop stateExpectObjectComma
|
||||
discard my.state.pop() # pop stateObject
|
||||
else:
|
||||
my.kind = jsonError
|
||||
my.err = errCurlyRiExpected
|
||||
of stateExpectColon:
|
||||
case tk
|
||||
of tkColon:
|
||||
my.state[i] = stateExpectValue
|
||||
next(my)
|
||||
else:
|
||||
my.kind = jsonError
|
||||
my.err = errColonExpected
|
||||
of stateExpectValue:
|
||||
case tk
|
||||
of tkString, tkInt, tkFloat, tkTrue, tkFalse, tkNull:
|
||||
my.state[i] = stateExpectObjectComma
|
||||
my.kind = JsonEventKind(ord(tk))
|
||||
of tkBracketLe:
|
||||
my.state[i] = stateExpectObjectComma
|
||||
my.state.add(stateArray)
|
||||
my.kind = jsonArrayStart
|
||||
of tkCurlyLe:
|
||||
my.state[i] = stateExpectObjectComma
|
||||
my.state.add(stateObject)
|
||||
my.kind = jsonObjectStart
|
||||
else:
|
||||
my.kind = jsonError
|
||||
my.err = errExprExpected
|
||||
|
||||
proc raiseParseErr*(p: JsonParser, msg: string) {.noinline, noreturn.} =
|
||||
## raises an `EJsonParsingError` exception.
|
||||
raise newException(JsonParsingError, errorMsgExpected(p, msg))
|
||||
|
||||
proc eat*(p: var JsonParser, tok: TokKind) =
|
||||
if p.tok == tok: discard getTok(p)
|
||||
else: raiseParseErr(p, tokToStr[tok])
|
||||
@@ -1826,8 +1826,8 @@ when isMainModule:
|
||||
doAssert(runeSubStr(s, 17, 1) == "€")
|
||||
# echo runeStrAtPos(s, 18) # index error
|
||||
|
||||
doAssert(runeSubStr(s, 0) == "Hänsel ««: 10,00€")
|
||||
doAssert(runeSubStr(s, -18) == "Hänsel ««: 10,00€")
|
||||
doAssert(runeSubStr(s, 0) == "Hänsel ««: 10,00€")
|
||||
doAssert(runeSubStr(s, -18) == "Hänsel ««: 10,00€")
|
||||
doAssert(runeSubStr(s, 10) == ": 10,00€")
|
||||
doAssert(runeSubStr(s, 18) == "")
|
||||
doAssert(runeSubStr(s, 0, 10) == "Hänsel ««")
|
||||
@@ -1840,7 +1840,7 @@ when isMainModule:
|
||||
doAssert(runeSubStr(s, -6, 5) == "10,00")
|
||||
doAssert(runeSubStr(s, -6, -1) == "10,00")
|
||||
|
||||
doAssert(runeSubStr(s, 0, 100) == "Hänsel ««: 10,00€")
|
||||
doAssert(runeSubStr(s, -100, 100) == "Hänsel ««: 10,00€")
|
||||
doAssert(runeSubStr(s, 0, 100) == "Hänsel ««: 10,00€")
|
||||
doAssert(runeSubStr(s, -100, 100) == "Hänsel ««: 10,00€")
|
||||
doAssert(runeSubStr(s, 0, -100) == "")
|
||||
doAssert(runeSubStr(s, 100, -100) == "")
|
||||
|
||||
145
lib/std/varints.nim
Normal file
145
lib/std/varints.nim
Normal file
@@ -0,0 +1,145 @@
|
||||
#
|
||||
#
|
||||
# Nim's Runtime Library
|
||||
# (c) Copyright 2018 Nim contributors
|
||||
#
|
||||
# See the file "copying.txt", included in this
|
||||
# distribution, for details about the copyright.
|
||||
#
|
||||
|
||||
## Note this API is still experimental! A variable length integer
|
||||
## encoding implementation inspired by SQLite.
|
||||
|
||||
const
|
||||
maxVarIntLen* = 9 ## the maximal number of bytes a varint can take
|
||||
|
||||
proc readVu64*(z: openArray[byte]; pResult: var uint64): int =
|
||||
if z[0] <= 240:
|
||||
pResult = z[0]
|
||||
return 1
|
||||
if z[0] <= 248:
|
||||
if z.len < 2: return 0
|
||||
pResult = (uint64 z[0] - 241) * 256 + uint64 z[1] + 240
|
||||
return 2
|
||||
if z.len < int(z[0]-246): return 0
|
||||
if z[0] == 249:
|
||||
pResult = 2288u64 + 256u64*z[1].uint64 + z[2].uint64
|
||||
return 3
|
||||
if z[0] == 250:
|
||||
pResult = (z[1].uint64 shl 16u64) + (z[2].uint64 shl 8u64) + z[3].uint64
|
||||
return 4
|
||||
let x = (z[1].uint64 shl 24) + (z[2].uint64 shl 16) + (z[3].uint64 shl 8) + z[4].uint64
|
||||
if z[0] == 251:
|
||||
pResult = x
|
||||
return 5
|
||||
if z[0] == 252:
|
||||
pResult = (((uint64)x) shl 8) + z[5].uint64
|
||||
return 6
|
||||
if z[0] == 253:
|
||||
pResult = (((uint64)x) shl 16) + (z[5].uint64 shl 8) + z[6].uint64
|
||||
return 7
|
||||
if z[0] == 254:
|
||||
pResult = (((uint64)x) shl 24) + (z[5].uint64 shl 16) + (z[6].uint64 shl 8) + z[7].uint64
|
||||
return 8
|
||||
pResult = (((uint64)x) shl 32) +
|
||||
(0xffffffff'u64 and ((z[5].uint64 shl 24) +
|
||||
(z[6].uint64 shl 16) + (z[7].uint64 shl 8) + z[8].uint64))
|
||||
return 9
|
||||
|
||||
proc varintWrite32(z: var openArray[byte]; y: uint32) =
|
||||
z[0] = uint8(y shr 24)
|
||||
z[1] = uint8(y shr 16)
|
||||
z[2] = uint8(y shr 8)
|
||||
z[3] = uint8(y)
|
||||
|
||||
proc writeVu64*(z: var openArray[byte], x: uint64): int =
|
||||
## Write a varint into z. The buffer z must be at least 9 characters
|
||||
## long to accommodate the largest possible varint. Returns the number of
|
||||
## bytes used.
|
||||
if x <= 240:
|
||||
z[0] = uint8 x
|
||||
return 1
|
||||
if x <= 2287:
|
||||
let y = uint32(x - 240)
|
||||
z[0] = uint8(y shr 8 + 241)
|
||||
z[1] = uint8(y and 255)
|
||||
return 2
|
||||
if x <= 67823:
|
||||
let y = uint32(x - 2288)
|
||||
z[0] = 249
|
||||
z[1] = uint8(y shr 8)
|
||||
z[2] = uint8(y and 255)
|
||||
return 3
|
||||
let y = uint32 x
|
||||
let w = uint32(x shr 32)
|
||||
if w == 0:
|
||||
if y <= 16777215:
|
||||
z[0] = 250
|
||||
z[1] = uint8(y shr 16)
|
||||
z[2] = uint8(y shr 8)
|
||||
z[3] = uint8(y)
|
||||
return 4
|
||||
z[0] = 251
|
||||
varintWrite32(toOpenArray(z, 1, z.high-1), y)
|
||||
return 5
|
||||
if w <= 255:
|
||||
z[0] = 252
|
||||
z[1] = uint8 w
|
||||
varintWrite32(toOpenArray(z, 2, z.high-2), y)
|
||||
return 6
|
||||
if w <= 65535:
|
||||
z[0] = 253
|
||||
z[1] = uint8(w shr 8)
|
||||
z[2] = uint8 w
|
||||
varintWrite32(toOpenArray(z, 3, z.high-3), y)
|
||||
return 7
|
||||
if w <= 16777215:
|
||||
z[0] = 254
|
||||
z[1] = uint8(w shr 16)
|
||||
z[2] = uint8(w shr 8)
|
||||
z[3] = uint8 w
|
||||
varintWrite32(toOpenArray(z, 4, z.high-4), y)
|
||||
return 8
|
||||
z[0] = 255
|
||||
varintWrite32(toOpenArray(z, 1, z.high-1), w)
|
||||
varintWrite32(toOpenArray(z, 5, z.high-5), y)
|
||||
return 9
|
||||
|
||||
proc sar(a, b: int64): int64 =
|
||||
{.emit: [result, " = ", a, " >> ", b, ";"].}
|
||||
|
||||
proc sal(a, b: int64): int64 =
|
||||
{.emit: [result, " = ", a, " << ", b, ";"].}
|
||||
|
||||
proc encodeZigzag*(x: int64): uint64 {.inline.} =
|
||||
uint64(sal(x, 1)) xor uint64(sar(x, 63))
|
||||
|
||||
proc decodeZigzag*(x: uint64): int64 {.inline.} =
|
||||
let casted = cast[int64](x)
|
||||
result = (`shr`(casted, 1)) xor (-(casted and 1))
|
||||
|
||||
when isMainModule:
|
||||
#import random
|
||||
|
||||
var dest: array[50, byte]
|
||||
var got: uint64
|
||||
|
||||
for test in [0xFFFF_FFFF_FFFFF_FFFFu64, 77u64, 0u64, 10_000_000u64, uint64(high(int64)),
|
||||
uint64(high(int32)),uint64(low(int32)),uint64(low(int64))]:
|
||||
let wrLen = writeVu64(dest, test)
|
||||
let rdLen = readVu64(dest, got)
|
||||
assert wrLen == rdLen
|
||||
echo(if got == test: "YES" else: "NO")
|
||||
echo "number is ", got
|
||||
|
||||
if encodeZigzag(decodeZigzag(test)) != test:
|
||||
echo "Failure for ", test, " ", encodeZigzag(decodeZigzag(test)), " ", decodeZigzag(test)
|
||||
|
||||
# check this also works for floats:
|
||||
for test in [0.0, 0.1, 2.0, +Inf, Nan, NegInf]:
|
||||
let t = cast[uint64](test)
|
||||
let wrLenB = writeVu64(dest, t)
|
||||
let rdLenB = readVu64(dest, got)
|
||||
assert wrLenB == rdLenB
|
||||
echo rdLenB
|
||||
echo(if cast[float64](got) == test: "YES" else: "NO")
|
||||
@@ -3808,7 +3808,9 @@ template doAssert*(cond: bool, msg = "") =
|
||||
bind instantiationInfo
|
||||
{.line: instantiationInfo().}:
|
||||
if not cond:
|
||||
raiseAssert(astToStr(cond) & ' ' & msg)
|
||||
raiseAssert(astToStr(cond) & ' ' &
|
||||
instantiationInfo(-1, false).fileName & '(' &
|
||||
$instantiationInfo(-1, false).line & ") " & msg)
|
||||
|
||||
iterator items*[T](a: seq[T]): T {.inline.} =
|
||||
## iterates over each item of `a`.
|
||||
|
||||
Reference in New Issue
Block a user