test-suite/tests/reader.test

   1 ;;;; reader.test --- Reader test.    -*- coding: iso-8859-1; mode: scheme -*-
   2 ;;;;
   3 ;;;; Copyright (C) 1999, 2001, 2002, 2003, 2007, 2008, 2009, 2010, 2011,
   4 ;;;;   2013 Free Software Foundation, Inc.
   5 ;;;; Jim Blandy <jimb@red-bean.com>
   6 ;;;;
   7 ;;;; This library is free software; you can redistribute it and/or
   8 ;;;; modify it under the terms of the GNU Lesser General Public
   9 ;;;; License as published by the Free Software Foundation; either
  10 ;;;; version 3 of the License, or (at your option) any later version.
  11 ;;;;
  12 ;;;; This library is distributed in the hope that it will be useful,
  13 ;;;; but WITHOUT ANY WARRANTY; without even the implied warranty of
  14 ;;;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  15 ;;;; Lesser General Public License for more details.
  16 ;;;;
  17 ;;;; You should have received a copy of the GNU Lesser General Public
  18 ;;;; License along with this library; if not, write to the Free Software
  19 ;;;; Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  20
  21 (define-module (test-suite reader)
  22   :use-module (srfi srfi-1)
  23   :use-module (test-suite lib))
  24
  25
  26 (define exception:eof
  27   (cons 'read-error "end of file$"))
  28 (define exception:unexpected-rparen
  29   (cons 'read-error "unexpected \")\"$"))
  30 (define exception:unexpected-rsqbracket
  31   (cons 'read-error "unexpected \"]\"$"))
  32 (define exception:unterminated-block-comment
  33   (cons 'read-error "unterminated `#. \\.\\.\\. .#' comment$"))
  34 (define exception:unknown-character-name
  35   (cons 'read-error "unknown character name .*$"))
  36 (define exception:unknown-sharp-object
  37   (cons 'read-error "Unknown # object: .*$"))
  38 (define exception:eof-in-string
  39   (cons 'read-error "end of file in string constant$"))
  40 (define exception:eof-in-symbol
  41   (cons 'read-error "end of file while reading symbol$"))
  42 (define exception:illegal-escape
  43   (cons 'read-error "illegal character in escape sequence: .*$"))
  44 (define exception:missing-expression
  45   (cons 'read-error "no expression after #;"))
  46 (define exception:mismatched-paren
  47   (cons 'read-error "mismatched close paren"))
  48
  49
  50 (define (read-string s)
  51   (with-input-from-string s (lambda () (read))))
  52
  53 (define (with-read-options opts thunk)
  54   (let ((saved-options (read-options)))
  55     (dynamic-wind
  56         (lambda ()
  57           (read-options opts))
  58         thunk
  59         (lambda ()
  60           (read-options saved-options)))))
  61
  62 \f
  63 (with-test-prefix "reading"
  64   (pass-if "0"
  65     (equal? (read-string "0") 0))
  66   (pass-if "1++i"
  67     (equal? (read-string "1++i") '1++i))
  68   (pass-if "1+i+i"
  69     (equal? (read-string "1+i+i") '1+i+i))
  70   (pass-if "1+e10000i"
  71     (equal? (read-string "1+e10000i") '1+e10000i))
  72   (pass-if "-nan.0-1i"
  73     (not (equal? (imag-part (read-string "-nan.0-1i"))
  74                  (imag-part (read-string "-nan.0+1i")))))
  75
  76   ;; At one time the arg list for "Unknown # object: ~S" didn't make it out
  77   ;; of read.c.  Check that `format' can be applied to this error.
  78   (pass-if "error message on bad #"
  79     (catch #t
  80            (lambda ()
  81              (read-string "#ZZZ")
  82              ;; oops, this # is supposed to be unrecognised
  83              #f)
  84            (lambda (key subr message args rest)
  85              (apply format #f message args)
  86              ;; message and args are ok
  87              #t)))
  88
  89   (pass-if "block comment"
  90     (equal? '(+ 1 2 3)
  91             (read-string "(+ 1 #! this is a\ncomment !# 2 3)")))
  92
  93   (pass-if "block comment finishing s-exp"
  94     (equal? '(+ 2)
  95             (read-string "(+ 2 #! a comment\n!#\n) ")))
  96
  97   (pass-if "R6RS lexeme comment"
  98     (equal? '(+ 1 2 3)
  99             (read-string "(+ 1 #!r6rs 2 3)")))
 100
 101   (pass-if "partial R6RS lexeme comment"
 102     (equal? '(+ 1 2 3)
 103             (read-string "(+ 1 #!r6r !# 2 3)")))
 104
 105   (pass-if "R6RS/SRFI-30 block comment"
 106     (equal? '(+ 1 2 3)
 107             (read-string "(+ 1 #| this is a\ncomment |# 2 3)")))
 108
 109   (pass-if "R6RS/SRFI-30 nested block comment"
 110     (equal? '(a b c)
 111             (read-string "(a b c #| d #| e |# f |#)")))
 112
 113   (pass-if "R6RS/SRFI-30 nested block comment (2)"
 114     (equal? '(a b c)
 115             (read-string "(a b c #|||||||#)")))
 116
 117   (pass-if "R6RS/SRFI-30 nested block comment (3)"
 118     (equal? '(a b c)
 119             (read-string "(a b c #||||||||#)")))
 120
 121   (pass-if "R6RS/SRFI-30 block comment syntax overridden"
 122     ;; To be compatible with 1.8 and earlier, we should be able to override
 123     ;; this syntax.
 124     (with-fluids ((%read-hash-procedures (fluid-ref %read-hash-procedures)))
 125       (read-hash-extend #\| (lambda args 'not))
 126       (fold (lambda (x y result)
 127               (and result (eq? x y)))
 128             #t
 129             (read-string "(this is #| a comment)")
 130             `(this is not a comment))))
 131
 132   (pass-if "unprintable symbol"
 133     ;; The reader tolerates unprintable characters for symbols.
 134     (equal? (string->symbol "\x01\x02\x03")
 135             (read-string "\x01\x02\x03")))
 136
 137   (pass-if "CR recognized as a token delimiter"
 138     ;; In 1.8.3, character 0x0d was not recognized as a delimiter.
 139     (equal? (read-string "one\x0dtwo") 'one))
 140
 141   (pass-if "returned strings are mutable"
 142     ;; Per R5RS Section 3.4, "Storage Model", `read' is supposed to return
 143     ;; mutable objects.
 144     (let ((str (with-input-from-string "\"hello, world\"" read)))
 145       (string-set! str 0 #\H)
 146       (string=? str "Hello, world")))
 147
 148   (pass-if "square brackets are parens"
 149     (equal? '() (read-string "[]")))
 150
 151   (pass-if-exception "paren mismatch" exception:unexpected-rparen
 152                      (read-string "'[)"))
 153
 154   (pass-if-exception "paren mismatch (2)" exception:unexpected-rsqbracket
 155                      (read-string "'(]"))
 156
 157   (pass-if-exception "paren mismatch (3)" exception:mismatched-paren
 158                      (read-string "'(foo bar]"))
 159
 160   (pass-if-exception "paren mismatch (4)" exception:mismatched-paren
 161                      (read-string "'[foo bar)")))
 162
 163
 164 \f
 165 (pass-if-exception "radix passed to number->string can't be zero"
 166   exception:out-of-range
 167   (number->string 10 0))
 168 (pass-if-exception "radix passed to number->string can't be one either"
 169   exception:out-of-range
 170   (number->string 10 1))
 171
 172 \f
 173 (with-test-prefix "mismatching parentheses"
 174   (pass-if-exception "opening parenthesis"
 175     exception:eof
 176     (read-string "("))
 177   (pass-if-exception "closing parenthesis following mismatched opening"
 178     exception:unexpected-rparen
 179     (read-string ")"))
 180   (pass-if-exception "closing square bracket following mismatched opening"
 181     exception:unexpected-rsqbracket
 182     (read-string "]"))
 183   (pass-if-exception "opening vector parenthesis"
 184     exception:eof
 185     (read-string "#("))
 186   (pass-if-exception "closing parenthesis following mismatched vector opening"
 187      exception:unexpected-rparen
 188      (read-string ")")))
 189
 190 \f
 191 (with-test-prefix "exceptions"
 192
 193   ;; Reader exceptions: although they are not documented, they may be relied
 194   ;; on by some programs, hence these tests.
 195
 196   (pass-if-exception "unterminated block comment"
 197     exception:unterminated-block-comment
 198     (read-string "(+ 1 #! comment\n..."))
 199   (pass-if-exception "R6RS/SRFI-30 unterminated nested block comment"
 200     exception:unterminated-block-comment
 201     (read-string "(foo #| bar #| |#)"))
 202   (pass-if-exception "unknown character name"
 203     exception:unknown-character-name
 204     (read-string "#\\theunknowncharacter"))
 205   (pass-if-exception "unknown sharp object"
 206     exception:unknown-sharp-object
 207     (read-string "#?"))
 208   (pass-if-exception "eof in string"
 209     exception:eof-in-string
 210     (read-string "\"the string that never ends"))
 211   (pass-if-exception "illegal escape in string"
 212     exception:illegal-escape
 213     (read-string "\"some string \\???\"")))
 214
 215 \f
 216 (with-test-prefix "read-options"
 217   (pass-if "case-sensitive"
 218     (not (eq? 'guile 'GuiLe)))
 219   (pass-if "case-insensitive"
 220     (eq? 'guile
 221          (with-read-options '(case-insensitive)
 222            (lambda ()
 223              (read-string "GuiLe")))))
 224   (pass-if "prefix keywords"
 225     (eq? #:keyword
 226          (with-read-options '(keywords prefix case-insensitive)
 227            (lambda ()
 228              (read-string ":KeyWord")))))
 229   (pass-if "prefix non-keywords"
 230     (symbol? (with-read-options '(keywords prefix)
 231                (lambda ()
 232                  (read-string "srfi88-keyword:")))))
 233   (pass-if "postfix keywords"
 234     (eq? #:keyword
 235          (with-read-options '(keywords postfix)
 236            (lambda ()
 237              (read-string "keyword:")))))
 238   (pass-if "long postfix keywords"
 239     (eq? #:keyword0123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789
 240          (with-read-options '(keywords postfix)
 241            (lambda ()
 242              (read-string "keyword0123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789:")))))
 243   (pass-if "`:' is not a postfix keyword (per SRFI-88)"
 244     (eq? ':
 245          (with-read-options '(keywords postfix)
 246            (lambda ()
 247              (read-string ":")))))
 248   (pass-if "no positions"
 249     (let ((sexp (with-read-options '()
 250                   (lambda ()
 251                     (read-string "(+ 1 2 3)")))))
 252       (and (not (source-property sexp 'line))
 253            (not (source-property sexp 'column)))))
 254   (pass-if "positions"
 255     (let ((sexp (with-read-options '(positions)
 256                   (lambda ()
 257                     (read-string "(+ 1 2 3)")))))
 258       (and (equal? (source-property sexp 'line) 0)
 259            (equal? (source-property sexp 'column) 0))))
 260   (pass-if "positions on quote"
 261     (let ((sexp (with-read-options '(positions)
 262                    (lambda ()
 263                     (read-string "'abcde")))))
 264       (and (equal? (source-property sexp 'line) 0)
 265            (equal? (source-property sexp 'column) 0))))
 266   (pass-if "position of SCSH block comment"
 267     ;; In Guile 2.0.0 the reader would not update the port's position
 268     ;; when reading an SCSH block comment.
 269     (let ((sexp (with-read-options '(positions)
 270                   (lambda ()
 271                     (read-string "#!foo\nbar\nbaz\n!#\n(hello world)\n")))))
 272       (= 4 (source-property sexp 'line))))
 273
 274   (with-test-prefix "r6rs-hex-escapes"
 275       (pass-if-exception "non-hex char in two-digit hex-escape"
 276       exception:illegal-escape
 277       (with-read-options '(r6rs-hex-escapes)
 278         (lambda ()
 279           (with-input-from-string "\"\\x0g;\"" read))))
 280
 281     (pass-if-exception "non-hex char in four-digit hex-escape"
 282       exception:illegal-escape
 283       (with-read-options '(r6rs-hex-escapes)
 284         (lambda ()
 285           (with-input-from-string "\"\\x000g;\"" read))))
 286
 287     (pass-if-exception "non-hex char in six-digit hex-escape"
 288       exception:illegal-escape
 289       (with-read-options '(r6rs-hex-escapes)
 290         (lambda ()
 291           (with-input-from-string "\"\\x00000g;\"" read))))
 292
 293     (pass-if-exception "no semicolon at termination of one-digit hex-escape"
 294       exception:illegal-escape
 295       (with-read-options '(r6rs-hex-escapes)
 296         (lambda ()
 297           (with-input-from-string "\"\\x0\"" read))))
 298
 299     (pass-if-exception "no semicolon at termination of three-digit hex-escape"
 300       exception:illegal-escape
 301       (with-read-options '(r6rs-hex-escapes)
 302         (lambda ()
 303           (with-input-from-string "\"\\x000\"" read))))
 304
 305     (pass-if "two-digit hex escape"
 306       (eqv?
 307        (with-read-options '(r6rs-hex-escapes)
 308          (lambda ()
 309            (string-ref (with-input-from-string "\"--\\xff;--\"" read) 2)))
 310        (integer->char #xff)))
 311
 312     (pass-if "four-digit hex escape"
 313       (eqv?
 314        (with-read-options '(r6rs-hex-escapes)
 315          (lambda ()
 316            (string-ref (with-input-from-string "\"--\\x0100;--\"" read) 2)))
 317        (integer->char #x0100)))
 318
 319     (pass-if "six-digit hex escape"
 320       (eqv?
 321        (with-read-options '(r6rs-hex-escapes)
 322          (lambda ()
 323            (string-ref (with-input-from-string "\"--\\x010300;--\"" read) 2)))
 324        (integer->char #x010300)))
 325
 326     (pass-if "escaped characters match non-escaped ASCII characters"
 327       (string=?
 328        (with-read-options '(r6rs-hex-escapes)
 329          (lambda ()
 330            (with-input-from-string "\"\\x41;\\x0042;\\x000043;\"" read)))
 331        "ABC"))
 332
 333     (pass-if "write R6RS string escapes"
 334        (let* ((s1 (apply string
 335                          (map integer->char '(#x8 ; backspace
 336                                               #x18 ; cancel
 337                                               #x20 ; space
 338                                               #x30 ; zero
 339                                               #x40 ; at sign
 340                                               ))))
 341               (s2 (with-read-options '(r6rs-hex-escapes)
 342                      (lambda ()
 343                       (with-output-to-string
 344                         (lambda () (write s1)))))))
 345          (lset= eqv?
 346                 (string->list s2)
 347                 (list #\" #\\ #\b #\\ #\x #\1 #\8 #\; #\space #\0 #\@ #\"))))
 348
 349     (pass-if "display R6RS string escapes"
 350       (string=?
 351        (with-read-options '(r6rs-hex-escapes)
 352          (lambda ()
 353            (let ((pt (open-output-string))
 354                  (s1 (apply string (map integer->char
 355                                         '(#xFF #x100 #xFFF #x1000 #xFFFF #x10000)))))
 356              (set-port-encoding! pt "ASCII")
 357              (set-port-conversion-strategy! pt 'escape)
 358              (display s1 pt)
 359              (get-output-string pt))))
 360        "\\xff;\\x100;\\xfff;\\x1000;\\xffff;\\x10000;"))
 361
 362     (pass-if "one-digit hex escape"
 363       (eqv? (with-input-from-string "#\\xA" read)
 364             (integer->char #x0A)))
 365
 366     (pass-if "two-digit hex escape"
 367       (eqv? (with-input-from-string "#\\xFF" read)
 368             (integer->char #xFF)))
 369
 370     (pass-if "four-digit hex escape"
 371       (eqv? (with-input-from-string "#\\x00FF" read)
 372             (integer->char #xFF)))
 373
 374     (pass-if "eight-digit hex escape"
 375       (eqv? (with-input-from-string "#\\x00006587" read)
 376             (integer->char #x6587)))
 377
 378     (pass-if "write R6RS escapes"
 379       (string=?
 380        (with-read-options '(r6rs-hex-escapes)
 381          (lambda ()
 382            (with-output-to-string
 383              (lambda ()
 384                (write (integer->char #x80))))))
 385        "#\\x80")))
 386
 387   (with-test-prefix "hungry escapes"
 388     (pass-if "default not hungry"
 389       ;; Assume default setting of not hungry.
 390       (equal? (with-input-from-string "\"foo\\\n  bar\""
 391                 read)
 392               "foo  bar"))
 393     (pass-if "hungry"
 394       (dynamic-wind
 395         (lambda ()
 396           (read-enable 'hungry-eol-escapes))
 397         (lambda ()
 398           (equal? (with-input-from-string "\"foo\\\n  bar\""
 399                     read)
 400                   "foobar"))
 401         (lambda ()
 402           (read-disable 'hungry-eol-escapes))))))
 403
 404 (with-test-prefix "per-port-read-options"
 405   (pass-if "case-sensitive"
 406     (equal? '(guile GuiLe gUIle)
 407             (with-read-options '(case-insensitive)
 408               (lambda ()
 409                 (with-input-from-string "GUIle #!no-fold-case GuiLe gUIle"
 410                   (lambda ()
 411                     (list (read) (read) (read))))))))
 412   (pass-if "case-insensitive"
 413     (equal? '(GUIle guile guile)
 414             (with-input-from-string "GUIle #!fold-case GuiLe gUIle"
 415               (lambda ()
 416                 (list (read) (read) (read)))))))
 417
 418 (with-test-prefix "#;"
 419   (for-each
 420    (lambda (pair)
 421      (pass-if (car pair)
 422        (equal? (with-input-from-string (car pair) read) (cdr pair))))
 423
 424    '(("#;foo 10". 10)
 425      ("#;(10 20 30) foo" . foo)
 426      ("#;   (10 20 30) foo" . foo)
 427      ("#;\n10\n20" . 20)))
 428
 429   (pass-if "#;foo"
 430     (eof-object? (with-input-from-string "#;foo" read)))
 431
 432   (pass-if-exception "#;"
 433     exception:missing-expression
 434     (with-input-from-string "#;" read))
 435   (pass-if-exception "#;("
 436     exception:eof
 437     (with-input-from-string "#;(" read)))
 438
 439 (with-test-prefix "#'"
 440   (for-each
 441    (lambda (pair)
 442      (pass-if (car pair)
 443        (equal? (with-input-from-string (car pair) read) (cdr pair))))
 444
 445    '(("#'foo". (syntax foo))
 446      ("#`foo" . (quasisyntax foo))
 447      ("#,foo" . (unsyntax foo))
 448      ("#,@foo" . (unsyntax-splicing foo)))))
 449
 450 (with-test-prefix "#{}#"
 451   (pass-if (equal? (read-string "#{}#") '#{}#))
 452   (pass-if (not (equal? (read-string "(a #{.}# b)") '(a . b))))
 453   (pass-if (equal? (read-string "#{a}#") 'a))
 454   (pass-if (equal? (read-string "#{a b}#") '#{a b}#))
 455   (pass-if-exception "#{" exception:eof-in-symbol
 456                      (read-string "#{"))
 457   (pass-if (equal? (read-string "#{a\\x20;b}#") '#{a b}#)))
 458
 459 (begin-deprecated
 460  (with-test-prefix "deprecated #{}# escapes"
 461    (pass-if (equal? (read-string "#{a\\ b}#") '#{a b}#))))
 462
 463 ;;; Local Variables:
 464 ;;; eval: (put 'with-read-options 'scheme-indent-function 1)
 465 ;;; End: