src/code/readtable.lisp

   1 ;;;; READTABLEs
   2
   3 ;;;; This software is part of the SBCL system. See the README file for
   4 ;;;; more information.
   5 ;;;;
   6 ;;;; This software is derived from the CMU CL system, which was
   7 ;;;; written at Carnegie Mellon University and released into the
   8 ;;;; public domain. The software is in the public domain and is
   9 ;;;; provided with absolutely no warranty. See the COPYING and CREDITS
  10 ;;;; files for more information.
  11
  12 (in-package "SB!IMPL")
  13
  14 (sb!xc:deftype attribute-table ()
  15   '(simple-array (unsigned-byte 8) (#.base-char-code-limit)))
  16
  17 ;;; constants for readtable character attributes. These are all as in
  18 ;;; the manual.
  19 ;;;
  20 ;;; FIXME: wait a minute.  Firstly, I doubt they're in the manual.
  21 ;;; Secondly, the numerical order of these constants is coupled with
  22 ;;; code in CHAR-CLASS{,2,3} in the reader implementation, so beware
  23 ;;; when changing them.
  24 (defconstant +char-attr-whitespace+ 0)
  25 (defconstant +char-attr-terminating-macro+ 1)
  26 (defconstant +char-attr-single-escape+ 2)
  27 (defconstant +char-attr-multiple-escape+ 3)
  28 (defconstant +char-attr-constituent+ 4)
  29 (defconstant +char-attr-constituent-dot+ 5)
  30 (defconstant +char-attr-constituent-expt+ 6)
  31 (defconstant +char-attr-constituent-slash+ 7)
  32 (defconstant +char-attr-constituent-digit+ 8)
  33 (defconstant +char-attr-constituent-sign+ 9)
  34 ;;; the following two are not static but depend on *READ-BASE*.
  35 ;;; DECIMAL-DIGIT is for characters being digits in base 10 but not in
  36 ;;; base *READ-BASE* (which is therefore perforce smaller than 10);
  37 ;;; DIGIT-OR-EXPT is for characters being both exponent markers and
  38 ;;; digits in base *READ-BASE* (which is therefore perforce larger
  39 ;;; than 10).  -- CSR, 2004-03-16
  40 (defconstant +char-attr-constituent-decimal-digit+ 10)
  41 (defconstant +char-attr-constituent-digit-or-expt+ 11)
  42
  43 (defconstant +char-attr-package-delimiter+ 12)
  44 (defconstant +char-attr-invalid+ 13)
  45 ;; Meta: there is no such function as READ-UNQUALIFIED-TOKEN. No biggie.
  46 (defconstant +char-attr-delimiter+ 14) ; (a fake for READ-UNQUALIFIED-TOKEN)
  47
  48 (sb!xc:defstruct (readtable (:conc-name nil)
  49                             (:constructor make-readtable ())
  50                             (:predicate readtablep)
  51                             ;; ANSI requires a CL:COPY-READTABLE to do
  52                             ;; a deep copy, so the DEFSTRUCT-generated
  53                             ;; default is not suitable.
  54                             (:copier nil))
  55   "A READTABLE is a data structure that maps characters into syntax
  56 types for the Common Lisp expression reader."
  57   ;; The CHARACTER-ATTRIBUTE-TABLE is a vector of BASE-CHAR-CODE-LIMIT
  58   ;; integers for describing the character type. Conceptually, there
  59   ;; are 4 distinct "primary" character attributes:
  60   ;; +CHAR-ATTR-WHITESPACE+, +CHAR-ATTR-TERMINATING-MACRO+,
  61   ;; +CHAR-ATTR-ESCAPE+, and +CHAR-ATTR-CONSTITUENT+. Non-terminating
  62   ;; macros (such as the symbol reader) have the attribute
  63   ;; +CHAR-ATTR-CONSTITUENT+.
  64   ;;
  65   ;; In order to make READ-TOKEN fast, all this information is stored
  66   ;; in the character attribute table by having different varieties of
  67   ;; constituents.
  68   (character-attribute-array
  69    (make-array base-char-code-limit
  70                :element-type '(unsigned-byte 8)
  71                :initial-element +char-attr-constituent+)
  72    :type attribute-table
  73    :read-only t)
  74   (character-attribute-hash-table (make-hash-table)
  75    :type hash-table
  76    :read-only t)
  77   ;; The CHARACTER-MACRO-TABLE is a vector of BASE-CHAR-CODE-LIMIT
  78   ;; functions. One of these functions called with appropriate
  79   ;; arguments whenever any non-WHITESPACE character is encountered
  80   ;; inside READ-PRESERVING-WHITESPACE. These functions are used to
  81   ;; implement user-defined read-macros, system read-macros, and the
  82   ;; number-symbol reader.
  83   (character-macro-array
  84    (make-array base-char-code-limit :initial-element nil)
  85    :type (simple-vector #.base-char-code-limit)
  86    :read-only t)
  87   (character-macro-hash-table (make-hash-table) :type hash-table
  88    :read-only t)
  89   (%readtable-case 0 :type (mod 4))
  90   ;; Element type to use when reading a string literal with no extended-chars.
  91   ;; The system itself prefers base-string, but otherwise it is a contentious
  92   ;; issue. We don't (by default) use base-strings, because people often write:
  93   ;;   (SETF (CHAR (READ-STRING S) 0) #\PILE_OF_POO),
  94   ;; or more likely, something the effect of which resembles
  95   ;;   (SETF (CHAR (ADJUST-ARRAY "" 10) 0) #\SMILE)
  96   ;; which are each dubious constructs, because they assume READ to produce
  97   ;; strings capable of holding any char. The latter further assumes something
  98   ;; about compilation, because in that example, considering that there are no
  99   ;; characters in the literal, it is unclear whether the array should
 100   ;; be similar-as-constant to an array of base-char or array of character.
 101   ;; While indeed SBCL prints base-strings readably (if *PRINT-READABLY* is T)
 102   ;; using #. syntax, the question is what the writer of the code intended
 103   ;; if (s)he did not know that the string should have been expressly
 104   ;; specified via #.(MAKE-STRING ... :ELEMENT-TYPE) or somesuch.
 105   (%readtable-string-preference 'base-char :type (member character base-char))
 106   ;; With symbols, it's fairly clear that immutability of print names
 107   ;; renders the distinction between the kinds of string in the symbol-name
 108   ;; as being less relevant. If you expect (copy-seq (string asymbol))
 109   ;; to produce a certain type of string, your code is unportable anyway.
 110   (%readtable-symbol-preference 'base-char :type (member character base-char))
 111   (%readtable-normalization #!+sb-unicode t #!-sb-unicode nil :type boolean))
 112
 113 (defconstant +readtable-upcase+ 0)
 114 (defconstant +readtable-downcase+ 1)
 115
 116 (declaim (freeze-type readtable))