ReadStream.st
author Claus Gittinger <cg@exept.de>
Sat, 11 Nov 1995 16:28:41 +0100
changeset 530 07d0bce293c9
parent 521 70533ec40482
child 546 4fb312758027
permissions -rw-r--r--
uff - version methods changed to return stings

"
 COPYRIGHT (c) 1988 by Claus Gittinger
	      All Rights Reserved

 This software is furnished under a license and may be used
 only in accordance with the terms of that license and with the
 inclusion of the above copyright notice.   This software may not
 be provided or otherwise made available to, or used by, any
 other person.  No title to or ownership of the software is
 hereby transferred.
"

PositionableStream subclass:#ReadStream
       instanceVariableNames:''
       classVariableNames:''
       poolDictionaries:''
       category:'Streams'
!

!ReadStream class methodsFor:'documentation'!

copyright
"
 COPYRIGHT (c) 1988 by Claus Gittinger
	      All Rights Reserved

 This software is furnished under a license and may be used
 only in accordance with the terms of that license and with the
 inclusion of the above copyright notice.   This software may not
 be provided or otherwise made available to, or used by, any
 other person.  No title to or ownership of the software is
 hereby transferred.
"
!

version
    ^ '$Header: /cvs/stx/stx/libbasic/ReadStream.st,v 1.18 1995-11-11 15:26:16 cg Exp $'
!

documentation
"
    ReadStream defines protocol for reading streamwise over collections. 
"
! !

!ReadStream methodsFor:'converting'!

readStream
    "return a readStream from the receiver. Since this is already
     a readStream, return self."

    ^ self
! !

!ReadStream methodsFor:'queries'!

isReadable
    ^ true
! !

!ReadStream methodsFor:'reading'!

peek
    "return the next element; do NOT advance read pointer.
     return nil, if there is no next element.
     - tuned for a bit more speed on String/ByteArray/Array-Streams"

%{  /* NOCONTEXT */

    REGISTER int pos;
    unsigned ch;
    OBJ coll;
    OBJ cls, p, l;
    extern OBJ ByteArray;

    coll = _INST(collection);
    p = _INST(position);
    l = _INST(readLimit);

    if (__isNonNilObject(coll) && __bothSmallInteger(p, l)) {

	pos = _intVal(p);
	if (pos > _intVal(l)) {
	    RETURN ( nil );
	}

	if (pos > 0) {
	    cls = __qClass(coll);
	    if (cls == String) {
		if (pos <= _stringSize(coll)) {
		    ch = _stringVal(coll)[pos-1];
		    RETURN ( _MKCHARACTER(ch) );
		}
	    } else if (cls == ByteArray) {
		if (pos <= _byteArraySize(coll)) {
		    ch = _ByteArrayInstPtr(coll)->ba_element[pos-1];
		    RETURN ( _MKSMALLINT(ch) );
		}
	    } else if (cls == Array) {
		if (pos <= _arraySize(coll)) {
		    RETURN ( _ArrayInstPtr(coll)->a_element[pos-1]);
		}
	    }
	}
    }
%}
.
    (position > readLimit) ifTrue:[^ nil].
    ^ collection at:position
!

next
    "return the next element; advance read pointer.
     return nil, if there is no next element.
     - tuned for a bit more speed on String/ByteArray/Array-Streams"

    |ret|

%{  /* NOCONTEXT */

    REGISTER int pos;
    unsigned ch;
    OBJ coll, p, l;

    coll = _INST(collection);
    p = _INST(position);
    l = _INST(readLimit);

    if (__isNonNilObject(coll) && __bothSmallInteger(p, l)) {

	pos = _intVal(p);
	if (pos > _intVal(l)) {
	    RETURN ( nil );
	}
	if (pos > 0) {
	    OBJ cls;

	    cls = __qClass(coll);
	    if (cls == String) {
		if (pos <= _stringSize(coll)) {
		    _INST(position) = _MKSMALLINT(pos + 1);
		    ch = _stringVal(coll)[pos-1];
		    RETURN ( _MKCHARACTER(ch) );
		}
	    } else if (cls == ByteArray) {
		if (pos <= _byteArraySize(coll)) {
		    _INST(position) = _MKSMALLINT(pos + 1);
		    ch = _ByteArrayInstPtr(coll)->ba_element[pos-1];
		    RETURN ( _MKSMALLINT(ch) );
		}
	    } else if (cls == Array) {
		if (pos <= _arraySize(coll)) {
		    _INST(position) = _MKSMALLINT(pos + 1);
		    RETURN ( _ArrayInstPtr(coll)->a_element[pos-1]);
		}
	    }
	}
    }
%}
.
    (position > readLimit) ifTrue:[^ nil].
    ret := collection at:position.
    position := position + 1.
    ^ ret
!

nextPeek
    "advance read pointer return the peek element.
     this is equivalent to (self next; peek).
     - tuned for speed on String-Streams for faster scanning"

%{  /* NOCONTEXT */
    OBJ coll, l, p;

    coll = _INST(collection);
    p = _INST(position);
    l = _INST(readLimit);

    if (__isString(coll) && __bothSmallInteger(p, l)) {
	REGISTER int pos;
	unsigned ch;

	pos = _intVal(p);
	if (pos > _intVal(l)) {
	    RETURN ( nil );
	}
	if ((pos > 0) && (pos <= _stringSize(coll))) {
	    _INST(position) = _MKSMALLINT(pos + 1);
	    pos = pos + 1;
	    if (pos <= _stringSize(coll)) {
		ch = _stringVal(coll)[pos-1];
		RETURN ( _MKCHARACTER(ch) );
	    }
	    RETURN ( nil );
	}
    }
%}
.
    (position > readLimit) ifTrue:[^ nil].
    position := position + 1.
    (position > readLimit) ifTrue:[^ nil].
    ^ collection at:position
!

nextDecimalInteger
    "read the next integer in radix 10. dont skip whitespace.
     - tuned for speed on String-Streams for faster scanning"

    |value nextOne|
%{
    int pos, limit, sz;
    REGISTER unsigned char *cp;
    REGISTER unsigned ch;
    int val = 0;
    OBJ coll, p, l;

    coll = _INST(collection);
    p = _INST(position);
    l = _INST(readLimit);
    
    if (__isString(coll) && __bothSmallInteger(p, l)) {

	pos = _intVal(p);
	limit = _intVal(l);
	sz = __qSize(coll) - OHDR_SIZE;
	if (sz < limit)
	    limit = sz; 
	cp = _stringVal(coll) + pos - 1;

	for (;;) {
	    if (pos > limit) break;
	    ch = *cp;

	    if ((ch < '0') || (ch > '9')) break;
	    val = val * 10 + (ch - '0');
	    pos++;
	    if (val > (_MAX_INT / 10)) goto oops;
	    cp++;
	}
	_INST(position) = _MKSMALLINT(pos);
	return _MKSMALLINT(val);
    }
oops:
    value = _MKSMALLINT(val);
%}
.
    "fall-back for non-string streams - we have to continue where
     above primitive left off, in case of a large integer ...
     (instead of doing a super nextDecimalInteger)"

    nextOne := self peek.
    [nextOne notNil and:[nextOne isDigitRadix:10]] whileTrue:[
	value = (value * 10) + nextOne digitValue.
	nextOne := self nextPeek
    ].
    ^ value
!

nextAlphaNumericWord
    "read the next word (i.e. up to non letter-or-digit).
     return a string containing those characters.
     - tuned for speed on String-Streams for faster scanning"
%{
    /* speedup, if collection is a string */

    int pos, limit, sz;
    int len;
    char buffer[256];
    REGISTER unsigned char *cp;
    REGISTER unsigned ch;
    OBJ coll, p, l;
    OBJ __MKSTRING_L();

    coll = _INST(collection);
    p = _INST(position);
    l = _INST(readLimit);
    
    if (__isString(coll) && __bothSmallInteger(p, l)) {

	pos = _intVal(p);
	limit = _intVal(l);
	sz = __qSize(coll) - OHDR_SIZE;
	if (sz < limit)
	    limit = sz; 
	cp = _stringVal(coll) + pos - 1;

	for (;;) {
	    if (pos > limit) break;
	    ch = *cp;

	    if (ch > ' ') break;
	    if ((ch != ' ') && (ch != '\t') && (ch != '\r')
	     && (ch != '\n') && (ch != 0x0b)) break;
	    cp++;
	    pos++;
	}

	len = 0;
	for (;;) {
	    if (pos > limit) break;
	    ch = *cp & 0xFF;

	    if (! (((ch >= 'a') && (ch <= 'z')) ||
		   ((ch >= 'A') && (ch <= 'Z')) ||
		   ((ch >= '0') && (ch <= '9'))))
		break;
	    buffer[len++] = ch;
	    if (len >= (sizeof(buffer)-1)) {
		/* emergency */
		break;
	    }
	    pos++;
	    cp++;
	}

	_INST(position) = _MKSMALLINT(pos);
	buffer[len] = '\0';
	RETURN ( (len != 0) ? __MKSTRING_L(buffer, len COMMA_CON) : nil );
    }
%}
.
    ^ super nextAlphaNumericWord
!

nextSymbol
    "read the next selector-symbol (i.e. up to non letter-or-digit).
     return a string containing those characters.
     - tuned for speed on String-Streams for faster scanning"
%{
    int pos, limit, sz;
    int len;
    char buffer[256];
    REGISTER unsigned char *cp;
    REGISTER unsigned ch;
    OBJ coll, p, l;
    OBJ __MKSTRING_L();

    coll = _INST(collection);
    p = _INST(position);
    l = _INST(readLimit);

    if (__isString(coll) && __bothSmallInteger(p, l)) {

	pos = _intVal(p);
	limit = _intVal(l);
	sz = __qSize(coll) - OHDR_SIZE;
	if (sz < limit)
	    limit = sz; 
	cp = _stringVal(coll) + pos - 1;

	len = 0;
	for (;;) {
	    if (pos > limit) break;
	    ch = *cp;

	    if (! (((ch >= 'a') && (ch <= 'z')) ||
		   ((ch >= 'A') && (ch <= 'Z')) ||
		   ((ch >= '0') && (ch <= '9')) ||
		   (ch == ':')))
		break;
	    buffer[len++] = ch;
	    if (len >= (sizeof(buffer)-1)) {
		/* emergency */
		break;
	    }
	    pos++;
	    cp++;
	}

	_INST(position) = _MKSMALLINT(pos);
	buffer[len] = '\0';
	RETURN ( (len != 0) ? __MKSTRING_L(buffer, len COMMA_CON) : nil );
    }
%}
.
    ^ super nextSymbol
!

skipThrough:anObject
    "skip all objects up-to and including anObject, return anObject on success,
     nil if end-of-stream is reached before. 
     On success, the next read operation will return the element after anObject.
     - reimplemented for speed on String-Streams for faster scanning"

%{  /* NOCONTEXT */
    OBJ coll, p, l;

    coll = _INST(collection);
    p = _INST(position);
    l = _INST(readLimit);

    if (__isString(coll)
     && __isCharacter(anObject)
     && __bothSmallInteger(p, l)) {
	REGISTER unsigned char *chars;
	REGISTER int pos, limit;
	unsigned ch;
	int sz;

	pos = _intVal(p);
	if (pos <= 0) {
	    RETURN ( nil );
	}

	limit = _intVal(l);
	sz = _stringSize(coll);
	if (limit > sz) limit = sz;

	chars = (unsigned char *)(_stringVal(coll) + pos - 1);
	ch = _intVal(_characterVal(anObject)) & 0xFF;
	while (pos < limit) {
	    if (*chars == ch) {
		ch = *++chars;
		pos++;
		_INST(position) = _MKSMALLINT(pos);
		RETURN ( anObject );
	    }
	    chars++;
	    pos++;
	}
	RETURN ( nil );
    }
%}
.
    ^ super skipThrough:anObject
!

skipSeparators
    "skip all whitespace; next will return next non-white-space element.
     Return the peeked character or nil, if the end-of-stream was reached.
     - reimplemented for speed on String-Streams for faster scanning"

%{  /* NOCONTEXT */
    OBJ coll, p, l;

    coll = _INST(collection);
    p = _INST(position);
    l = _INST(readLimit);

    if (__isString(coll) && __bothSmallInteger(p, l)) {
	REGISTER unsigned char *chars;
	REGISTER unsigned ch;
	REGISTER int pos;
	int limit;

	pos = _intVal(p);
	if (pos <= 0) {
	    RETURN ( nil );
	}

	limit = _intVal(l);
	if (limit > (__qSize(coll) - OHDR_SIZE))
	    limit = __qSize(coll) - OHDR_SIZE;

	chars = (unsigned char *)(_stringVal(coll) + pos - 1);
	while (pos <= limit) {
	    ch = *chars++;
	    if ((ch != ' ') && (ch != '\t') && (ch != '\r')
	     && (ch != '\n') && (ch != 0x0B)) {
		_INST(position) = _MKSMALLINT(pos);
		RETURN ( _MKCHARACTER(ch) );
	    }
	    pos++;
	}
	_INST(position) = _MKSMALLINT(pos);
	RETURN ( nil );
    }
%}
.
    ^ super skipSeparators
!

skipSeparatorsExceptCR
    "skip all whitespace except newlines;
     next will return next non-white-space element.
     - reimplemented for speed on String-Streams for faster scanning"

%{  /* NOCONTEXT */

    OBJ coll, p, l;

    coll = _INST(collection);
    p = _INST(position);
    l = _INST(readLimit);

    if (__isString(coll) && __bothSmallInteger(p, l)) {
	REGISTER unsigned char *chars;
	REGISTER unsigned ch;
	REGISTER int pos;
	int limit;

	pos = _intVal(p);
	if (pos <= 0) {
	    RETURN ( nil );
	}

	limit = _intVal(l);
	if (limit > (__qSize(coll) - OHDR_SIZE))
	    limit = __qSize(coll) - OHDR_SIZE;

	chars = (unsigned char *)(_stringVal(coll) + pos - 1);
	while (pos <= limit) {
	    ch = *chars++;
	    if ((ch != ' ') && (ch != '\t') && (ch != '\r') && (ch != 0x0B)) {
		_INST(position) = _MKSMALLINT(pos);
		RETURN ( _MKCHARACTER(ch) );
	    }
	    pos++;
	}
	_INST(position) = _MKSMALLINT(pos);
	RETURN ( nil );
    }
%}
.
    ^ super skipSeparatorsExceptCR
!

skipToAll:aCollection
    "skip for the sequence given by the argument, aCollection;
     return nil if not found, self otherwise. On a successful match, next read
     will return elements of aCollection."

    |oldPos buffer l first idx|

    oldPos := self position.
    l := aCollection size.
    first := aCollection at:1.
    [self atEnd] whileFalse:[
	buffer := self next:l.
	buffer = aCollection ifTrue:[
	    self position:(self position - l).
	    ^ self
	].
	idx := buffer indexOf:first startingAt:2.
	idx == 0 ifFalse:[
	    self position:(self position - l + idx - 1)
	]
    ].
    self position:oldPos.
    ^ nil

    "
     |s|
     s := ReadStream on:'12345678901234567890'.
     s skipToAll:'901'.
     s next:4
    "
! !

!ReadStream methodsFor:'writing'!

nextPut:anElement
    "catch write access to readstreams - report an error"

    self shouldNotImplement
! !