/*
Returns zero if matching rule not found or exception occured.
*/
static const UCell* _parseBlock( UThread* ut, BlockParser* pe,
const UCell* rit, const UCell* rend,
UIndex* spos )
{
const UCell* tval;
int32_t repMin;
int32_t repMax;
UAtom atom;
const UBuffer* iblk = pe->blk;
UIndex pos = *spos;
match:
while( rit != rend )
{
switch( ur_type(rit) )
{
case UT_WORD:
atom = ur_atom(rit);
if( atom < UT_BI_COUNT )
{
// Datatype
if( pos >= pe->inputEnd )
goto failed;
tval = iblk->ptr.cell + pos;
if( ur_type(tval) != atom )
{
/*
if( atom == UT_NUMBER )
{
if( ur_is(tval,UT_INT) || ur_is(tval,UT_DECIMAL) )
goto type_matched;
}
*/
goto failed;
}
//type_matched:
++rit;
++pos;
}
else switch( atom )
{
case UR_ATOM_OPT:
++rit;
repMin = 0;
repMax = 1;
goto repeat;
case UR_ATOM_ANY:
++rit;
repMin = 0;
repMax = 0x7fffffff;
goto repeat;
case UR_ATOM_SOME:
++rit;
repMin = 1;
repMax = 0x7fffffff;
goto repeat;
case UR_ATOM_BREAK:
pe->exception = PARSE_EX_BREAK;
*spos = pos;
return 0;
case UR_ATOM_BAR:
goto complete;
case UR_ATOM_TO:
case UR_ATOM_THRU:
{
const UCell* ci;
const UCell* ce;
UAtom ratom = ur_atom(rit);
++rit;
if( rit == rend )
return 0;
ci = iblk->ptr.cell + pos;
ce = iblk->ptr.cell + pe->inputEnd;
if( ur_is(rit, UT_WORD) )
{
if( ur_atom(rit) < UT_BI_COUNT )
{
atom = ur_atom(rit);
while( ci != ce )
{
if( ur_type(ci) == atom )
break;
++ci;
}
if( ci == ce )
goto failed;
pos = ci - iblk->ptr.cell;
if( ratom == UR_ATOM_THRU )
++pos;
++rit;
break;
}
else
{
tval = ur_wordCell( ut, rit );
CHECK_WORD( tval )
}
}
else
{
tval = rit;
}
if( ur_is(tval, UT_BLOCK) )
{
// TODO: If block then all values must match.
BLK_RULE_ERROR( "to/thru block! not implemented" );
}
else
{
while( ci != ce )
{
if( ur_equal(ut, ci, tval) )
break;
++ci;
}
if( ci == ce )
goto failed;
pos = ci - iblk->ptr.cell;
if( ratom == UR_ATOM_THRU )
++pos;
}
++rit;
}
break;
case UR_ATOM_INTO:
++rit;
if( rit == rend || ! ur_is(rit, UT_BLOCK) )
{
BLK_RULE_ERROR( "parse into expected block" );
}
tval = iblk->ptr.cell + pos;
if( ! ur_is(tval, UT_BLOCK) )
goto failed;
if( ur_isShared( tval->series.buf ) )
goto failed;
{
BlockParser ip;
UBlockIter bi;
UIndex parsePos = 0;
ip.eval = pe->eval;
ip.blk = ur_bufferSer( tval );
ip.inputBuf = tval->series.buf;
ip.inputEnd = ip.blk->used;
ip.sliced = 0;
ip.exception = PARSE_EX_NONE;
ur_blkSlice( ut, &bi, rit );
tval = _parseBlock( ut, &ip, bi.it, bi.end, &parsePos );
iblk = _acquireInput( ut, pe );
if( ! tval )
{
if( ip.exception == PARSE_EX_ERROR )
{
pe->exception = PARSE_EX_ERROR;
ur_appendTrace( ut, rit->series.buf, 0 );
return 0;
}
if( ip.exception != PARSE_EX_BREAK )
goto failed;
}
}
++rit;
++pos;
break;
case UR_ATOM_SKIP:
repMin = 1;
skip:
if( (pos + repMin) > pe->inputEnd )
goto failed;
pos += repMin;
++rit;
break;
case UR_ATOM_SET:
++rit;
if( rit == rend )
goto unexpected_end;
if( ! ur_is(rit, UT_WORD) )
{
BLK_RULE_ERROR( "parse set expected word" );
}
{
UCell* cell = ur_wordCellM( ut, rit );
CHECK_WORD( cell )
*cell = iblk->ptr.cell[ pos ];
}
++rit;
break;
case UR_ATOM_PLACE:
++rit;
if( (rit != rend) && ur_is(rit, UT_WORD) )
{
tval = ur_wordCell( ut, rit++ );
CHECK_WORD( tval )
if( ur_is(tval, UT_BLOCK) )
{
pos = tval->series.it;
break;
}
}
BLK_RULE_ERROR( "place expected series word" );
//case UR_ATOM_COPY:
default:
{
tval = ur_wordCell( ut, rit );
CHECK_WORD( tval )
if( ur_is(tval, UT_BLOCK) )
{
goto match_block;
}
else
{
BLK_RULE_ERROR( "parse expected block" );
}
}
break;
}
break;
case UT_SETWORD:
{
UCell* cell = ur_wordCellM( ut, rit );
CHECK_WORD( cell )
++rit;
ur_setId( cell, UT_BLOCK );
ur_setSlice( cell, pe->inputBuf, pos, pe->inputEnd );
}
break;
case UT_GETWORD:
{
UCell* cell = ur_wordCellM( ut, rit );
CHECK_WORD( cell )
++rit;
if( ur_is(cell, UT_BLOCK) &&
(cell->series.buf == pe->inputBuf) )
cell->series.end = pos;
}
break;
case UT_LITWORD:
if( pos >= pe->inputEnd )
goto failed;
tval = iblk->ptr.cell + pos;
if( (ur_is(tval, UT_WORD) || ur_is(tval, UT_LITWORD))
&& (ur_atom(tval) == ur_atom(rit)) )
{
++rit;
++pos;
}
else
goto failed;
break;
case UT_INT:
repMin = ur_int(rit);
++rit;
if( rit == rend )
return 0;
if( ur_is(rit, UT_INT) )
{
repMax = ur_int(rit);
++rit;
}
else if( ur_is(rit, UT_WORD) && ur_atom(rit) == UR_ATOM_SKIP )
{
goto skip;
}
else
{
repMax = repMin;
}
goto repeat;
case UT_DATATYPE:
if( pos >= pe->inputEnd )
goto failed;
if( ! ur_isDatatype( iblk->ptr.cell + pos, rit ) )
goto failed;
++rit;
++pos;
break;
case UT_CHAR:
case UT_BINARY:
case UT_STRING:
case UT_FILE:
if( pos >= pe->inputEnd )
goto failed;
if( ! ur_equal( ut, iblk->ptr.cell + pos, rit ) )
goto failed;
++rit;
++pos;
break;
case UT_BLOCK:
tval = rit;
match_block:
{
UBlockIter bi;
UIndex rblkN = tval->series.buf;
ur_blkSlice( ut, &bi, tval );
tval = _parseBlock( ut, pe, bi.it, bi.end, &pos );
iblk = pe->blk;
if( ! tval )
{
if( pe->exception == PARSE_EX_ERROR )
{
ur_appendTrace( ut, rblkN, 0 );
return 0;
}
if( pe->exception == PARSE_EX_BREAK )
pe->exception = PARSE_EX_NONE;
else
goto failed;
}
}
++rit;
break;
case UT_PAREN:
if( UR_OK != pe->eval( ut, rit ) )
goto parse_err;
iblk = _acquireInput( ut, pe );
++rit;
break;
default:
BLK_RULE_ERROR( "invalid parse value" );
}
/*
Returns zero if matching rule not found or exception occured.
*/
static const UCell* _parseBin( UThread* ut, BinaryParser* pe,
const UCell* rit, const UCell* rend,
UIndex* spos )
{
const UCell* set = 0;
const UCell* tval;
uint32_t bitCount;
uint32_t field;
UBuffer* ibin = ur_buffer( pe->inputBufN );
uint8_t* in = ibin->ptr.b + *spos;
uint8_t* inEnd = ibin->ptr.b + pe->inputEnd;
match:
while( rit != rend )
{
switch( ur_type(rit) )
{
case UT_INT:
bitCount = ur_int(rit);
if( bitCount < 1 || bitCount > 32 )
{
ur_error( PARSE_ERR, "bit-field size must be 1 to 32" );
goto parse_err;
}
if( bitCount > 24 )
{
uint32_t high;
in = pullBits( pe, bitCount - 16, in, inEnd, &high );
if( ! in )
goto failed;
in = pullBits( pe, 16, in, inEnd, &field );
if( ! in )
goto failed;
field |= high << 16;
}
else
{
in = pullBits( pe, bitCount, in, inEnd, &field );
if( ! in )
goto failed;
}
goto set_field;
case UT_WORD:
switch( ur_atom(rit) )
{
case UR_ATOM_U8:
if( in == inEnd )
goto failed;
field = *in++;
goto set_field;
case UR_ATOM_U16:
if( (inEnd - in) < 2 )
goto failed;
if( pe->bigEndian )
field = (in[0] << 8) | in[1];
else
field = (in[1] << 8) | in[0];
in += 2;
goto set_field;
case UR_ATOM_U32:
if( (inEnd - in) < 4 )
goto failed;
if( pe->bigEndian )
field = (in[0] << 24) | (in[1] << 16) |
(in[2] << 8) | in[3];
else
field = (in[3] << 24) | (in[2] << 16) |
(in[1] << 8) | in[0];
in += 4;
goto set_field;
case UR_ATOM_SKIP:
++rit;
++in;
break;
#if 0
case UR_ATOM_MARK:
break;
case UR_ATOM_PLACE:
++rit;
if( (rit != rend) && ur_is(rit, UT_WORD) )
{
tval = ur_wordCell( ut, rit++ );
CHECK_WORD(tval);
if( ur_is(tval, UT_BINARY) )
{
pos = tval->series.it;
break;
}
}
ur_error( PARSE_ERR, "place expected series word" );
goto parse_err;
#endif
case UR_ATOM_COPY: // copy dest size
// word! int!/word!
++rit;
if( (rit != rend) && ur_is(rit, UT_WORD) )
{
UCell* res = ur_wordCellM( ut, rit );
CHECK_WORD(res);
if( ++rit != rend )
{
tval = rit++;
if( ur_is(tval, UT_WORD) )
{
tval = ur_wordCell( ut, tval );
CHECK_WORD(tval);
}
if( ur_is(tval, UT_INT) )
{
UBuffer* cb;
int size = ur_int(tval);
cb = ur_makeBinaryCell( ut, size, res );
cb->used = size;
memCpy( cb->ptr.b, in, size );
in += size;
break;
}
}
ur_error( PARSE_ERR, "copy expected int! count" );
goto parse_err;
}
ur_error( PARSE_ERR, "copy expected word! destination" );
goto parse_err;
case UR_ATOM_BIG_ENDIAN:
++rit;
pe->bigEndian = 1;
break;
case UR_ATOM_LITTLE_ENDIAN:
++rit;
pe->bigEndian = 0;
break;
default:
tval = ur_wordCell( ut, rit );
CHECK_WORD(tval);
if( ur_is(tval, UT_CHAR) )
goto match_char;
else if( ur_is(tval, UT_STRING) )
goto match_string;
else if( ur_is(tval, UT_BLOCK) )
goto match_block;
/*
else if( ur_is(tval, UT_BITSET) )
goto match_bitset;
*/
else
{
ur_error( PARSE_ERR,
"parse expected char!/string!/block!" );
goto parse_err;
}
break;
}
break;
case UT_SETWORD:
set = rit++;
while( (rit != rend) && ur_is(rit, UT_SETWORD) )
++rit;
break;
#if 0
case UT_GETWORD:
break;
case UT_INT:
repMin = ur_int(rit);
++rit;
if( rit == rend )
return 0;
if( ur_is(rit, UT_INT) )
{
repMax = ur_int(rit);
++rit;
}
else
{
repMax = repMin;
}
goto repeat;
#endif
case UT_CHAR:
match_char:
if( *in != ur_int(rit) )
goto failed;
++in;
++rit;
break;
case UT_BLOCK:
tval = rit;
match_block:
{
UBlockIter bi;
UIndex pos = in - ibin->ptr.b;
UIndex rblkN = tval->series.buf;
ur_blkSlice( ut, &bi, tval );
tval = _parseBin( ut, pe, bi.it, bi.end, &pos );
ibin = ur_buffer( pe->inputBufN );
if( ! tval )
{
if( pe->exception == PARSE_EX_ERROR )
{
ur_appendTrace( ut, rblkN, 0 );
return 0;
}
if( pe->exception == PARSE_EX_BREAK )
pe->exception = PARSE_EX_NONE;
else
goto failed;
}
in = ibin->ptr.b + pos;
inEnd = ibin->ptr.b + pe->inputEnd;
++rit;
}
break;
case UT_PAREN:
{
UIndex pos = in - ibin->ptr.b;
if( UR_OK != pe->eval( ut, rit ) )
goto parse_err;
/* Re-acquire pointer & check if input modified. */
ibin = ur_buffer( pe->inputBufN );
if( pe->sliced )
{
// We have no way to track changes to the end of a slice,
// so just make sure we remain in valid memery.
if( ibin->used < pe->inputEnd )
pe->inputEnd = ibin->used;
}
else
{
// Not sliced, track input end.
if( ibin->used != pe->inputEnd )
pe->inputEnd = ibin->used;
}
in = ibin->ptr.b + pos;
inEnd = ibin->ptr.b + pe->inputEnd;
++rit;
}
break;
case UT_STRING:
tval = rit;
match_string:
{
UBinaryIter bi;
int size;
ur_binSlice( ut, &bi, tval );
if( ur_strIsUcs2(bi.buf) )
goto bad_enc;
size = bi.end - bi.it;
if( size > (inEnd - in) )
goto failed;
if( match_pattern_8(in, inEnd, bi.it, bi.end) == bi.end )
{
in += size;
++rit;
}
else
goto failed;
}
break;
#if 0
case UT_BITSET:
tval = rit;
match_bitset:
if( pos >= pe->inputEnd )
goto failed;
{
const UBuffer* bin = ur_bufferSer( tval );
int c = istr->ptr.c[ pos ];
if( bitIsSet( bin->ptr.b, c ) )
{
++rit;
++pos;
}
else
goto failed;
}
break;
#endif
default:
ur_error( PARSE_ERR, "invalid parse value" );
//orDatatypeName( ur_type(rit) ) );
goto parse_err;
}
}
//complete:
*spos = in - ibin->ptr.b;
return rit;
set_field:
if( set )
{
UCell* val;
while( set != rit )
{
val = ur_wordCellM( ut, set++ );
CHECK_WORD(val);
ur_setId(val, UT_INT);
ur_int(val) = field;
}
set = 0;
}
++rit;
goto match;
failed:
*spos = in - ibin->ptr.b;
return 0;
bad_enc:
ur_error( ut, UR_ERR_INTERNAL,
"parse binary does not handle UCS2 strings" );
//goto parse_err;
parse_err:
pe->exception = PARSE_EX_ERROR;
return 0;
}
