os/persistentdata/persistentstorage/dbms/usql/UQ_LEXER.CPP
author sl
Tue, 10 Jun 2014 14:32:02 +0200
changeset 1 260cb5ec6c19
permissions -rw-r--r--
Update contrib.
sl@0
     1
// Copyright (c) 1998-2009 Nokia Corporation and/or its subsidiary(-ies).
sl@0
     2
// All rights reserved.
sl@0
     3
// This component and the accompanying materials are made available
sl@0
     4
// under the terms of "Eclipse Public License v1.0"
sl@0
     5
// which accompanies this distribution, and is available
sl@0
     6
// at the URL "http://www.eclipse.org/legal/epl-v10.html".
sl@0
     7
//
sl@0
     8
// Initial Contributors:
sl@0
     9
// Nokia Corporation - initial contribution.
sl@0
    10
//
sl@0
    11
// Contributors:
sl@0
    12
//
sl@0
    13
// Description:
sl@0
    14
// SQL parser tokeniser
sl@0
    15
// 
sl@0
    16
//
sl@0
    17
sl@0
    18
#include "UQ_STD.H"
sl@0
    19
sl@0
    20
// optimised tables for ASCII character set
sl@0
    21
sl@0
    22
const TUint8 KAlpha=0x1;
sl@0
    23
const TUint8 KDigitOr_=0x2;
sl@0
    24
sl@0
    25
const TUint8 KCharAttrib[]=
sl@0
    26
	{
sl@0
    27
	KDigitOr_,KDigitOr_,KDigitOr_,KDigitOr_,KDigitOr_,KDigitOr_,KDigitOr_,KDigitOr_,
sl@0
    28
	KDigitOr_,KDigitOr_,0,0,0,0,0,0,
sl@0
    29
	0,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,
sl@0
    30
	KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,
sl@0
    31
	KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,
sl@0
    32
	KAlpha,KAlpha,KAlpha,0,0,0,0,KDigitOr_,
sl@0
    33
	0,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,
sl@0
    34
	KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,
sl@0
    35
	KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,KAlpha,
sl@0
    36
	KAlpha,KAlpha,KAlpha
sl@0
    37
	};
sl@0
    38
sl@0
    39
#define ISALPHA(aChar) (TUint(aChar-'0')<=TUint('z'-'0') && KCharAttrib[aChar-'0']&KAlpha)
sl@0
    40
#define ISALPHA_DIGIT_OR_(aChar) (TUint(aChar-'0')<=TUint('z'-'0') && KCharAttrib[aChar-'0']&(KAlpha|KDigitOr_))
sl@0
    41
#define LCASE(aChar) (aChar|0x20)
sl@0
    42
sl@0
    43
// The keywords
sl@0
    44
// These are always stored as ASCII because DBMS has its own 
sl@0
    45
sl@0
    46
const TInt KMaxKeywordLength=9;
sl@0
    47
sl@0
    48
static const TText8 KSqlKeywords[][KMaxKeywordLength+1]=
sl@0
    49
	{
sl@0
    50
#define KEYWORD(s) #s
sl@0
    51
#include "UQ_KEYWD.H"
sl@0
    52
#undef KEYWORD
sl@0
    53
	};
sl@0
    54
const TInt KSqlKeywordCount=sizeof(KSqlKeywords)/sizeof(*KSqlKeywords);
sl@0
    55
sl@0
    56
#if defined(_ASSERTIONS)
sl@0
    57
TInt CheckKeywords()
sl@0
    58
//
sl@0
    59
// ensure that the keyword table is in alphabetical order
sl@0
    60
//
sl@0
    61
	{
sl@0
    62
	for (TInt ii=1;ii<KSqlKeywordCount;++ii)
sl@0
    63
		__ASSERT(TPtrC8(KSqlKeywords[ii-1])<TPtrC8(KSqlKeywords[ii]));
sl@0
    64
	return 1;
sl@0
    65
	}
sl@0
    66
#endif
sl@0
    67
sl@0
    68
// class TSqlLexer
sl@0
    69
sl@0
    70
TInt TSqlLexer::CompareKeyword(TInt aKeyword,const RSqlLiteral& aIdentifier)
sl@0
    71
//
sl@0
    72
// Check if the identifer in aIdentifier is a keyword
sl@0
    73
// uses a case-insensitive match, not folding
sl@0
    74
//
sl@0
    75
	{
sl@0
    76
	__ASSERT(TUint(aKeyword)<TUint(KSqlKeywordCount));
sl@0
    77
//
sl@0
    78
	const TText* ptr=aIdentifier.Ptr();
sl@0
    79
	const TText* end=aIdentifier.End();
sl@0
    80
	const TText8* pk=&KSqlKeywords[aKeyword][0];
sl@0
    81
	for (;;)
sl@0
    82
		{
sl@0
    83
		TUint ck=*pk++;
sl@0
    84
		if (ptr==end)
sl@0
    85
			return ck;
sl@0
    86
		if (!ck)
sl@0
    87
			return -1;
sl@0
    88
		TInt d=ck-LCASE(TUint(*ptr++));
sl@0
    89
		if (d)
sl@0
    90
			return d;
sl@0
    91
		}
sl@0
    92
	}
sl@0
    93
sl@0
    94
TSqlKeyword TSqlLexer::Keyword(const TSqlToken& aToken)
sl@0
    95
//
sl@0
    96
// non-member function: Return the keyword value
sl@0
    97
//
sl@0
    98
	{
sl@0
    99
	if (aToken==ESqlIdentifier)
sl@0
   100
		{
sl@0
   101
		TInt r=KSqlKeywordCount;
sl@0
   102
		TInt l=0;
sl@0
   103
		while (r>l)
sl@0
   104
			{
sl@0
   105
			TInt m=(l+r)>>1;
sl@0
   106
			TInt k=CompareKeyword(m,aToken.Literal());
sl@0
   107
			if (k>0)
sl@0
   108
				r=m;
sl@0
   109
			else if (k<0)
sl@0
   110
				l=m+1;
sl@0
   111
			else
sl@0
   112
				return TSqlKeyword(m);		// keyword
sl@0
   113
			}
sl@0
   114
		}
sl@0
   115
	// identifier
sl@0
   116
	return ESqlNotKeyword;
sl@0
   117
	}
sl@0
   118
sl@0
   119
TSqlLexer::TSqlLexer(const TDesC& aSql)
sl@0
   120
	: iNext(aSql.Ptr()),iEnd(iNext+aSql.Length())
sl@0
   121
	{
sl@0
   122
	__ASSERT(CheckKeywords());
sl@0
   123
	}
sl@0
   124
sl@0
   125
TSqlTokenType TSqlLexer::GetIdentifier(TSqlToken& aToken)
sl@0
   126
//
sl@0
   127
// Get a keyword or identifier. Do not resolve a keyword at this stage
sl@0
   128
//
sl@0
   129
	{
sl@0
   130
	const TText* end=iEnd;
sl@0
   131
	const TText* next=iNext-1;
sl@0
   132
	while (++next<end)
sl@0
   133
		{
sl@0
   134
		TUint ch=*next;
sl@0
   135
		if (ISALPHA_DIGIT_OR_(ch))
sl@0
   136
			continue;
sl@0
   137
		if (!TChar(ch).IsAlphaDigit())
sl@0
   138
			break;
sl@0
   139
		}
sl@0
   140
	aToken.iLiteral.SetText(iNext-1,next);
sl@0
   141
	iNext=next;
sl@0
   142
	return ESqlIdentifier;
sl@0
   143
	}
sl@0
   144
sl@0
   145
TInt TSqlLexer::GetInteger(TInt64 &aVal)
sl@0
   146
//
sl@0
   147
// A rather more optimised version of TLex::Val(TInt64&)
sl@0
   148
// initially accumulate the value in a TUint32, and only switch to 64-bit arithmetic
sl@0
   149
// if the value overflows. Always return a 64-bit value
sl@0
   150
//
sl@0
   151
	{
sl@0
   152
	const TUint KPreMultiplyLimit32=429496728;	//(KMaxTUint-9)/10
sl@0
   153
	const TUint KPreMultiplyLimit64=214748364;	//(KMaxTInt+1)/10
sl@0
   154
//
sl@0
   155
	const TText* ptr=iNext;
sl@0
   156
	const TText* const end=iEnd;
sl@0
   157
	__ASSERT(ptr<end);
sl@0
   158
	TUint sign=0;
sl@0
   159
	TUint c=*ptr;
sl@0
   160
	if (c=='-')
sl@0
   161
		{
sl@0
   162
		sign=1;
sl@0
   163
		if (++ptr==end)
sl@0
   164
			return KErrGeneral;
sl@0
   165
		c=*ptr;
sl@0
   166
		}
sl@0
   167
	else if (c=='+')
sl@0
   168
		{
sl@0
   169
		if (++ptr==end)
sl@0
   170
			return KErrGeneral;
sl@0
   171
		c=*ptr;
sl@0
   172
		}
sl@0
   173
	c-='0';
sl@0
   174
	if (c>=10u)
sl@0
   175
		return KErrGeneral;		// no digits at all
sl@0
   176
	TUint val32=c;
sl@0
   177
	while (++ptr<end)
sl@0
   178
		{
sl@0
   179
		c=*ptr-'0';
sl@0
   180
		if (c>=10u)
sl@0
   181
			break;
sl@0
   182
		if (val32>KPreMultiplyLimit32)
sl@0
   183
			goto overflow64;	// will not fit into 32 bit arithmetic
sl@0
   184
		val32*=10;
sl@0
   185
		val32+=c;
sl@0
   186
		}
sl@0
   187
// we have result, just set the sign and finish
sl@0
   188
	aVal=val32;
sl@0
   189
	goto checksign;
sl@0
   190
//
sl@0
   191
// continue the accumulation with a 64-bit integer
sl@0
   192
overflow64:
sl@0
   193
	aVal=val32;
sl@0
   194
	for (;;)
sl@0
   195
		{
sl@0
   196
		I64MUL10(aVal);
sl@0
   197
		aVal+=c;
sl@0
   198
		if (++ptr==end)
sl@0
   199
			break;
sl@0
   200
		c=*ptr-'0';
sl@0
   201
		if (c>=10u)
sl@0
   202
			break;
sl@0
   203
		if (I64HIGH(aVal)>KPreMultiplyLimit64)
sl@0
   204
			return KErrOverflow;	// the value is certain to overflow
sl@0
   205
		}
sl@0
   206
	if (I64HIGH(aVal)&0x80000000u)
sl@0
   207
		{	// greater than the "half way mark"
sl@0
   208
		if (!sign)
sl@0
   209
			return KErrOverflow;
sl@0
   210
		if (I64LOW(aVal)!=0)
sl@0
   211
			return KErrOverflow;
sl@0
   212
		}
sl@0
   213
checksign:
sl@0
   214
	iNext=ptr;
sl@0
   215
	if (sign)
sl@0
   216
		aVal=-aVal;
sl@0
   217
	return KErrNone;
sl@0
   218
	}
sl@0
   219
sl@0
   220
TSqlTokenType TSqlLexer::GetNumber(TSqlToken& aToken)
sl@0
   221
	{
sl@0
   222
	const TText* mark=--iNext;			// rewind past initial character
sl@0
   223
	// attempt to parse a integer
sl@0
   224
	TInt r=GetInteger(aToken.iLiteral.SetInt());
sl@0
   225
	if (r==KErrNone)
sl@0
   226
		{
sl@0
   227
		if (iNext<iEnd)
sl@0
   228
			{
sl@0
   229
			TUint c=*iNext;
sl@0
   230
			if (c!='.' && c!='e' && c!='E')
sl@0
   231
				return ESqlLiteralInt;		// it is an integer
sl@0
   232
			}
sl@0
   233
		else
sl@0
   234
			return ESqlLiteralInt;		// it is an integer
sl@0
   235
		}
sl@0
   236
	TLex lex(TPtrC(mark,iEnd-mark));
sl@0
   237
	r=lex.Val(aToken.iLiteral.SetReal(),TChar('.'));
sl@0
   238
	if (r!=KErrNone)
sl@0
   239
		return SqlError(r);
sl@0
   240
	iNext=mark+lex.Offset();
sl@0
   241
	return ESqlLiteralReal;
sl@0
   242
	}
sl@0
   243
sl@0
   244
TSqlTokenType TSqlLexer::GetString(TSqlToken& aToken)
sl@0
   245
	{
sl@0
   246
	const TText* next=iNext;
sl@0
   247
	const TText* end=iEnd;
sl@0
   248
	for (;;)
sl@0
   249
		{
sl@0
   250
		if (next==end)
sl@0
   251
			return SqlError();
sl@0
   252
		TUint c=*next++;
sl@0
   253
		if (c=='\'')
sl@0
   254
			{
sl@0
   255
			if (next==end)
sl@0
   256
				break;
sl@0
   257
			if (*next!='\'')
sl@0
   258
				break;
sl@0
   259
			next++;
sl@0
   260
			}
sl@0
   261
		}
sl@0
   262
	aToken.iLiteral.SetText(iNext,next-1);
sl@0
   263
	iNext=next;
sl@0
   264
	return ESqlLiteralText;
sl@0
   265
	}
sl@0
   266
sl@0
   267
TSqlTokenType TSqlLexer::GetDate(TSqlToken& aToken)
sl@0
   268
	{
sl@0
   269
	const TText* end=iEnd;
sl@0
   270
	const TText* next=iNext;
sl@0
   271
	do
sl@0
   272
		{
sl@0
   273
		if (next==end)
sl@0
   274
			return SqlError();
sl@0
   275
		} while (*next++!='#');
sl@0
   276
	TInt r=aToken.iLiteral.SetTime().Parse(TPtrC(iNext,(next-1)-iNext));
sl@0
   277
	if (r<0)
sl@0
   278
		return SqlError(r);
sl@0
   279
	iNext=next;
sl@0
   280
	return ESqlLiteralTime;
sl@0
   281
	}
sl@0
   282
sl@0
   283
TSqlTokenType TSqlLexer::GetNextToken(TSqlToken& aToken)
sl@0
   284
	{
sl@0
   285
	const TText* ptr=iNext;
sl@0
   286
	const TText* const end=iEnd;
sl@0
   287
	for (;;)
sl@0
   288
		{
sl@0
   289
		if (ptr==end)
sl@0
   290
			return ESqlEos;
sl@0
   291
		TUint ch=*ptr++;
sl@0
   292
		iNext=ptr;
sl@0
   293
		switch (ch)
sl@0
   294
			{
sl@0
   295
		case ' ':			// a "normal" space
sl@0
   296
			continue;
sl@0
   297
		case '0': case '1': case '2': case '3': case '4':	// literal number
sl@0
   298
		case '5': case '6': case '7': case '8': case '9':
sl@0
   299
		case '+': case '-': case '.':
sl@0
   300
			return GetNumber(aToken);
sl@0
   301
		case '\'':
sl@0
   302
			return GetString(aToken);
sl@0
   303
		case '#':
sl@0
   304
			return GetDate(aToken);
sl@0
   305
		case '*':
sl@0
   306
			return ESqlAsterisk;
sl@0
   307
		case ',':
sl@0
   308
			return ESqlComma;
sl@0
   309
		case '(':
sl@0
   310
			return ESqlLeftBracket;
sl@0
   311
		case ')':
sl@0
   312
			return ESqlRightBracket;
sl@0
   313
		case '=':
sl@0
   314
			return ESqlEqual;
sl@0
   315
		case '<':
sl@0
   316
			{
sl@0
   317
			ch=*ptr++;				
sl@0
   318
			if (ch=='=')
sl@0
   319
				{
sl@0
   320
				iNext=ptr;
sl@0
   321
				return ESqlLessEqual;
sl@0
   322
				}
sl@0
   323
			if (ch=='>')
sl@0
   324
				{
sl@0
   325
				iNext=ptr;
sl@0
   326
				return ESqlNotEqual;
sl@0
   327
				}
sl@0
   328
			return ESqlLess;
sl@0
   329
			}
sl@0
   330
		case '>':
sl@0
   331
			{
sl@0
   332
			ch=*ptr++;			
sl@0
   333
			if (ch=='=')
sl@0
   334
				{
sl@0
   335
				iNext=ptr;
sl@0
   336
				return ESqlGreaterEqual;
sl@0
   337
				}
sl@0
   338
			return ESqlGreater;
sl@0
   339
			}
sl@0
   340
		default:
sl@0
   341
			break;
sl@0
   342
			}
sl@0
   343
		if (ISALPHA(ch))
sl@0
   344
			return GetIdentifier(aToken);		// keyword or identifier
sl@0
   345
		TChar cc(ch);
sl@0
   346
		if (cc.IsAlpha())
sl@0
   347
			return GetIdentifier(aToken);		// keyword or identifier
sl@0
   348
		if (!cc.IsSpace())
sl@0
   349
			return SqlError();
sl@0
   350
		}
sl@0
   351
	}
sl@0
   352
	
sl@0
   353
const TText* TSqlLexer::Next() const
sl@0
   354
	{
sl@0
   355
	return iNext;
sl@0
   356
	}
sl@0
   357
void TSqlLexer::Set(const TText* aNext) 
sl@0
   358
	{
sl@0
   359
	iNext = aNext ;
sl@0
   360
	}