1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305
|
#
#
# The Nim Compiler
# (c) Copyright 2018 Andreas Rumpf
#
# See the file "copying.txt", included in this
# distribution, for details about the copyright.
#
## Layouter for nimpretty.
import idents, lexer, lineinfos, llstream, options, msgs, strutils,
pathutils
from os import changeFileExt
const
MaxLineLen = 80
LineCommentColumn = 30
type
SplitKind = enum
splitComma, splitParLe, splitAnd, splitOr, splitIn, splitBinary
SemicolonKind = enum
detectSemicolonKind, useSemicolon, dontTouch
Emitter* = object
config: ConfigRef
fid: FileIndex
lastTok: TTokType
inquote, lastTokWasTerse: bool
semicolons: SemicolonKind
col, lastLineNumber, lineSpan, indentLevel, indWidth*: int
keepIndents*: int
doIndentMore*: int
content: string
indentStack: seq[int]
fixedUntil: int # marks where we must not go in the content
altSplitPos: array[SplitKind, int] # alternative split positions
proc openEmitter*(em: var Emitter, cache: IdentCache;
config: ConfigRef, fileIdx: FileIndex) =
let fullPath = Absolutefile config.toFullPath(fileIdx)
if em.indWidth == 0:
em.indWidth = getIndentWidth(fileIdx, llStreamOpen(fullPath, fmRead),
cache, config)
if em.indWidth == 0: em.indWidth = 2
em.config = config
em.fid = fileIdx
em.lastTok = tkInvalid
em.inquote = false
em.col = 0
em.content = newStringOfCap(16_000)
em.indentStack = newSeqOfCap[int](30)
em.indentStack.add 0
em.lastLineNumber = 1
proc closeEmitter*(em: var Emitter) =
if fileExists(em.config.outFile) and readFile(em.config.outFile.string) == em.content:
discard "do nothing, see #9499"
return
var f = llStreamOpen(em.config.outFile, fmWrite)
if f == nil:
rawMessage(em.config, errGenerated, "cannot open file: " & em.config.outFile.string)
return
f.llStreamWrite em.content
llStreamClose(f)
proc countNewlines(s: string): int =
result = 0
for i in 0..<s.len:
if s[i] == '\L': inc result
proc calcCol(em: var Emitter; s: string) =
var i = s.len-1
em.col = 0
while i >= 0 and s[i] != '\L':
dec i
inc em.col
template wr(x) =
em.content.add x
inc em.col, x.len
template goodCol(col): bool = col in 40..MaxLineLen
const
openPars = {tkParLe, tkParDotLe,
tkBracketLe, tkBracketLeColon, tkCurlyDotLe,
tkCurlyLe}
splitters = openPars + {tkComma, tkSemicolon}
oprSet = {tkOpr, tkDiv, tkMod, tkShl, tkShr, tkIn, tkNotin, tkIs,
tkIsnot, tkNot, tkOf, tkAs, tkDotDot, tkAnd, tkOr, tkXor}
template rememberSplit(kind) =
if goodCol(em.col):
em.altSplitPos[kind] = em.content.len
template moreIndent(em): int =
(if em.doIndentMore > 0: em.indWidth*2 else: em.indWidth)
proc softLinebreak(em: var Emitter, lit: string) =
# XXX Use an algorithm that is outlined here:
# https://llvm.org/devmtg/2013-04/jasper-slides.pdf
# +2 because we blindly assume a comma or ' &' might follow
if not em.inquote and em.col+lit.len+2 >= MaxLineLen:
if em.lastTok in splitters:
while em.content.len > 0 and em.content[em.content.high] == ' ':
setLen(em.content, em.content.len-1)
wr("\L")
em.col = 0
for i in 1..em.indentLevel+moreIndent(em): wr(" ")
else:
# search backwards for a good split position:
for a in mitems(em.altSplitPos):
if a > em.fixedUntil:
var spaces = 0
while a+spaces < em.content.len and em.content[a+spaces] == ' ':
inc spaces
if spaces > 0: delete(em.content, a, a+spaces-1)
em.col = em.content.len - a
let ws = "\L" & repeat(' ', em.indentLevel+moreIndent(em))
em.content.insert(ws, a)
a = -1
break
proc emitTok*(em: var Emitter; L: TLexer; tok: TToken) =
template endsInWhite(em): bool =
em.content.len == 0 or em.content[em.content.high] in {' ', '\L'}
template endsInAlpha(em): bool =
em.content.len > 0 and em.content[em.content.high] in SymChars+{'_'}
proc emitComment(em: var Emitter; tok: TToken) =
let lit = strip fileSection(em.config, em.fid, tok.commentOffsetA, tok.commentOffsetB)
em.lineSpan = countNewlines(lit)
if em.lineSpan > 0: calcCol(em, lit)
if not endsInWhite(em):
wr(" ")
if em.lineSpan == 0 and max(em.col, LineCommentColumn) + lit.len <= MaxLineLen:
for i in 1 .. LineCommentColumn - em.col: wr(" ")
wr lit
if tok.tokType == tkComment and tok.literal.startsWith("#!nimpretty"):
case tok.literal
of "#!nimpretty off":
inc em.keepIndents
wr("\L")
em.lastLineNumber = tok.line + 1
of "#!nimpretty on":
dec em.keepIndents
em.lastLineNumber = tok.line
wr("\L")
#for i in 1 .. tok.indent: wr " "
wr tok.literal
em.col = 0
em.lineSpan = 0
return
var preventComment = false
if tok.tokType == tkComment and tok.line == em.lastLineNumber and tok.indent >= 0:
# we have an inline comment so handle it before the indentation token:
emitComment(em, tok)
preventComment = true
em.fixedUntil = em.content.high
elif tok.indent >= 0:
if em.lastTok in (splitters + oprSet) or em.keepIndents > 0:
em.indentLevel = tok.indent
else:
if tok.indent > em.indentStack[^1]:
em.indentStack.add tok.indent
else:
# dedent?
while em.indentStack.len > 1 and em.indentStack[^1] > tok.indent:
discard em.indentStack.pop()
em.indentLevel = em.indentStack.high * em.indWidth
#[ we only correct the indentation if it is not in an expression context,
so that code like
const splitters = {tkComma, tkSemicolon, tkParLe, tkParDotLe,
tkBracketLe, tkBracketLeColon, tkCurlyDotLe,
tkCurlyLe}
is not touched.
]#
# remove trailing whitespace:
while em.content.len > 0 and em.content[em.content.high] == ' ':
setLen(em.content, em.content.len-1)
wr("\L")
for i in 2..tok.line - em.lastLineNumber: wr("\L")
em.col = 0
for i in 1..em.indentLevel:
wr(" ")
em.fixedUntil = em.content.high
var lastTokWasTerse = false
case tok.tokType
of tokKeywordLow..tokKeywordHigh:
if endsInAlpha(em):
wr(" ")
elif not em.inquote and not endsInWhite(em) and
em.lastTok notin openPars and not em.lastTokWasTerse:
#and tok.tokType in oprSet
wr(" ")
if not em.inquote:
wr(TokTypeToStr[tok.tokType])
case tok.tokType
of tkAnd: rememberSplit(splitAnd)
of tkOr: rememberSplit(splitOr)
of tkIn, tkNotin:
rememberSplit(splitIn)
wr(" ")
else: discard
else:
# keywords in backticks are not normalized:
wr(tok.ident.s)
of tkColon:
wr(TokTypeToStr[tok.tokType])
wr(" ")
of tkSemicolon, tkComma:
wr(TokTypeToStr[tok.tokType])
rememberSplit(splitComma)
wr(" ")
of tkParDotLe, tkParLe, tkBracketDotLe, tkBracketLe,
tkCurlyLe, tkCurlyDotLe, tkBracketLeColon:
if tok.strongSpaceA > 0 and not em.endsInWhite:
wr(" ")
wr(TokTypeToStr[tok.tokType])
rememberSplit(splitParLe)
of tkParRi,
tkBracketRi, tkCurlyRi,
tkBracketDotRi,
tkCurlyDotRi,
tkParDotRi,
tkColonColon:
wr(TokTypeToStr[tok.tokType])
of tkDot:
lastTokWasTerse = true
wr(TokTypeToStr[tok.tokType])
of tkEquals:
if not em.inquote and not em.endsInWhite: wr(" ")
wr(TokTypeToStr[tok.tokType])
if not em.inquote: wr(" ")
of tkOpr, tkDotDot:
if (tok.strongSpaceA == 0 and tok.strongSpaceB == 0) or em.inquote:
# bug #9504: remember to not spacify a keyword:
lastTokWasTerse = true
# if not surrounded by whitespace, don't produce any whitespace either:
wr(tok.ident.s)
else:
if not em.endsInWhite: wr(" ")
wr(tok.ident.s)
template isUnary(tok): bool =
tok.strongSpaceB == 0 and tok.strongSpaceA > 0
if not isUnary(tok):
rememberSplit(splitBinary)
wr(" ")
of tkAccent:
if not em.inquote and endsInAlpha(em): wr(" ")
wr(TokTypeToStr[tok.tokType])
em.inquote = not em.inquote
of tkComment:
if not preventComment:
emitComment(em, tok)
of tkIntLit..tkStrLit, tkRStrLit, tkTripleStrLit, tkGStrLit, tkGTripleStrLit, tkCharLit:
let lit = fileSection(em.config, em.fid, tok.offsetA, tok.offsetB)
softLinebreak(em, lit)
if endsInAlpha(em) and tok.tokType notin {tkGStrLit, tkGTripleStrLit}: wr(" ")
em.lineSpan = countNewlines(lit)
if em.lineSpan > 0: calcCol(em, lit)
wr lit
of tkEof: discard
else:
let lit = if tok.ident != nil: tok.ident.s else: tok.literal
softLinebreak(em, lit)
if endsInAlpha(em): wr(" ")
wr lit
em.lastTok = tok.tokType
em.lastTokWasTerse = lastTokWasTerse
em.lastLineNumber = tok.line + em.lineSpan
em.lineSpan = 0
proc starWasExportMarker*(em: var Emitter) =
if em.content.endsWith(" * "):
setLen(em.content, em.content.len-3)
em.content.add("*")
dec em.col, 2
proc commaWasSemicolon*(em: var Emitter) =
if em.semicolons == detectSemicolonKind:
em.semicolons = if em.content.endsWith(", "): dontTouch else: useSemicolon
if em.semicolons == useSemicolon and em.content.endsWith(", "):
setLen(em.content, em.content.len-2)
em.content.add("; ")
proc curlyRiWasPragma*(em: var Emitter) =
if em.content.endsWith("}"):
setLen(em.content, em.content.len-1)
em.content.add(".}")
|