source: proto/SymbolTable/parabix2_symtab_pbs_log.py @ 1228

Last change on this file since 1228 was 1228, checked in by vla24, 8 years ago

Integrated symbol table with xmlwf. There are various implementations for the symbol table, please read /proto/SymbolTable/README_SymbolTable for more information.

File size: 21.1 KB
Line 
1# -*- coding: utf-8 -*-
2#
3# parabix2_compilable.py
4#
5# Parallel XML Parsing with Bitstream Addition
6#
7# - Complete prototype for all bitstream computations in Parabix2
8# - Optimized for compilation
9# - Separate compilation
10
11# Robert D. Cameron
12# July 29, 2010
13#
14
15#import bitutil
16
17class u8 ():
18  unibyte = 0
19  prefix = 0
20  prefix2 = 0
21  prefix3 = 0
22  prefix4 = 0
23  suffix = 0
24  badprefix = 0
25  xE0 = 0
26  xED = 0
27  xF0 = 0
28  xF4 = 0
29  xA0_xBF = 0
30  x80_x9F = 0
31  x90_xBF = 0
32  x80_x8F = 0
33  xEF = 0
34  xBF = 0
35  xBE = 0
36  scope22 = 0
37  scope32 = 0
38  scope33 = 0
39  scope42 = 0
40  scope43 = 0
41  scope44 = 0
42  xE0_scope = 0
43  xED_scope = 0
44  xF0_scope = 0
45  xF4_scope = 0
46  xEF_scope = 0
47 
48  FFFE_FFFF = 0
49  error = 0
50
51class Lex ():
52        CR = 0
53        LF = 0
54        HT = 0
55        SP = 0
56        CRLF = 0
57        RefStart = 0
58        Semicolon = 0 
59        Colon = 0
60        LAngle = 0
61        RAngle = 0
62        LBracket = 0
63        RBracket = 0
64        Exclam = 0
65        QMark = 0
66        Hyphen = 0
67        Equals = 0
68        SQuote = 0
69        DQuote = 0
70        Slash = 0
71        Hash = 0
72        x = 0
73        ASCII_name_start = 0
74        ASCII_name_char = 0
75        NameScan = 0
76        Digit = 0
77        Hex = 0
78        WS = 0
79        error = 0
80
81class Scope1 ():
82        RefStart = 0
83        LAngle = 0
84        Hyphen = 0
85        QMark = 0
86        RBracket = 0
87
88class CtCDPI_Callouts():
89        CD_end = 0
90        Ct_starts = 0
91        Ct_ends = 0
92        CD_starts = 0
93        CD_ends = 0
94        PI_starts = 0
95        PI_name_starts = 0
96        PI_name_ends = 0
97        PI_ends = 0
98        CtCDPI_mask = 0
99        error = 0
100
101class Ref_Callouts():
102        GenRef_starts = 0
103        GenRef_ends = 0
104        DecRef_starts = 0
105        DecRef_ends = 0
106        HexRef_starts = 0
107        HexRef_ends = 0
108        error = 0
109
110class Hash_data():
111        Hash_value = 0
112
113class Tag_Callouts():
114        ElemName_starts = 0
115        ElemName_ends = 0
116        ElemName_ends_1 = 0
117        ElemName_ends_2 = 0
118        ElemName_ends_3_to_4 = 0
119        ElemName_ends_5_to_8 = 0
120        ElemName_ends_9_to_16 = 0
121        ElemName_ends_17_and_longer = 0
122        AttName_starts = 0
123        AttName_ends = 0
124        AttVal_starts = 0
125        AttVal_ends = 0
126        AttVal_spans = 0
127        EmptyTag_marks = 0
128        EndTag_marks = 0
129        LAngleFollow = 0
130        error = 0
131
132class Basis_bits():     
133        bit_0 = 0
134        bit_1 = 0
135        bit_2 = 0
136        bit_3 = 0
137        bit_4 = 0
138        bit_5 = 0
139        bit_6 = 0
140        bit_7 = 0
141       
142class Check_streams():
143        misc_mask = 0
144        non_ascii_name_starts = 0
145        non_ascii_names = 0
146        tag_marks = 0 
147        name_follows = 0 
148        att_refs = 0 
149        error_mask = 0
150
151class Xml_names():
152        namespace_error = 0
153
154def Classify_bytes_Validate_utf8(basis_bits, lex, u8): 
155        temp1 = (basis_bits.bit_0 | basis_bits.bit_1);
156        temp2 = (basis_bits.bit_2 &~ basis_bits.bit_3);
157        temp3 = (temp2 &~ temp1);
158        temp4 = (basis_bits.bit_5 &~ basis_bits.bit_4);
159        temp5 = (basis_bits.bit_6 &~ basis_bits.bit_7);
160        temp6 = (temp4 & temp5);
161        lex.RefStart = (temp3 & temp6);
162        temp7 = (basis_bits.bit_2 & basis_bits.bit_3);
163        temp8 = (temp7 &~ temp1);
164        temp9 = (basis_bits.bit_4 &~ basis_bits.bit_5);
165        temp10 = (basis_bits.bit_6 & basis_bits.bit_7);
166        temp11 = (temp9 & temp10);
167        lex.Semicolon = (temp8 & temp11);
168        temp12 = (basis_bits.bit_4 & basis_bits.bit_5);
169        temp13 = (basis_bits.bit_6 | basis_bits.bit_7);
170        temp14 = (temp12 &~ temp13);
171        lex.LAngle = (temp8 & temp14);
172        temp15 = (temp12 & temp5);
173        lex.RAngle = (temp8 & temp15);
174        temp16 = (basis_bits.bit_1 &~ basis_bits.bit_0);
175        temp17 = (basis_bits.bit_3 &~ basis_bits.bit_2);
176        temp18 = (temp16 & temp17);
177        lex.LBracket = (temp18 & temp11);
178        temp19 = (basis_bits.bit_7 &~ basis_bits.bit_6);
179        temp20 = (temp12 & temp19);
180        lex.RBracket = (temp18 & temp20);
181        temp21 = (basis_bits.bit_4 | basis_bits.bit_5);
182        temp22 = (temp19 &~ temp21);
183        lex.Exclam = (temp3 & temp22);
184        temp23 = (temp12 & temp10);
185        lex.QMark = (temp8 & temp23);
186        lex.Hyphen = (temp3 & temp20);
187        lex.Equals = (temp8 & temp20);
188        temp24 = (temp4 & temp10);
189        lex.SQuote = (temp3 & temp24);
190        temp25 = (temp5 &~ temp21);
191        lex.DQuote = (temp3 & temp25);
192        lex.Slash = (temp3 & temp23);
193        temp26 = (temp10 &~ temp21);
194        lex.Hash = (temp3 & temp26);
195        temp27 = (temp16 & temp7);
196        temp28 = (temp9 &~ temp13);
197        lex.x = (temp27 & temp28);
198        temp29 = (temp9 & temp5);
199        lex.Colon = (temp8 & temp29);
200        temp30 = (temp18 & temp23);
201        temp31 = (temp30 | lex.Colon);
202        temp32 = (temp16 &~ basis_bits.bit_2);
203        temp33 = (basis_bits.bit_5 | temp10);
204        temp34 = (basis_bits.bit_4 & temp33);
205        temp35 = (~temp34);
206        temp36 = (temp21 | temp13);
207        temp37 = ((basis_bits.bit_3 & temp35)|(~(basis_bits.bit_3) & temp36));
208        temp38 = (temp32 & temp37);
209        temp39 = (temp31 | temp38);
210        temp40 = (temp16 & basis_bits.bit_2);
211        temp41 = (temp40 & temp37);
212        lex.ASCII_name_start = (temp39 | temp41);
213        temp42 = (temp30 | lex.Hyphen);
214        temp43 = (temp3 & temp15);
215        temp44 = (temp42 | temp43);
216        temp45 = (temp8 &~ temp34);
217        temp46 = (temp44 | temp45);
218        temp47 = (temp46 | temp38);
219        lex.ASCII_name_char = (temp47 | temp41);
220        lex.NameScan = (lex.ASCII_name_char | basis_bits.bit_0);
221        temp48 = (temp1 | basis_bits.bit_2);
222        x00_x1F = (~temp48);
223        temp49 = (basis_bits.bit_2 | basis_bits.bit_3);
224        temp50 = (temp1 | temp49);
225        lex.CR = (temp20 &~ temp50);
226        lex.LF = (temp29 &~ temp50);
227        temp51 = (temp9 & temp19);
228        lex.HT = (temp51 &~ temp50);
229        lex.SP = (temp3 &~ temp36);
230        temp52 = (temp20 | temp29);
231        temp53 = (temp52 | temp51);
232        temp54 = (temp53 &~ temp50);
233        lex.WS = (temp54 | lex.SP);
234        temp55 = (basis_bits.bit_5 | basis_bits.bit_6);
235        temp56 = (basis_bits.bit_4 & temp55);
236        lex.Digit = (temp8 &~ temp56);
237        temp57 = (temp16 &~ temp49);
238        temp58 = (temp57 &~ basis_bits.bit_4);
239        temp59 = (~temp10);
240        temp60 = ((basis_bits.bit_5 & temp59)|(~(basis_bits.bit_5) & temp13));
241        temp61 = (temp58 & temp60);
242        temp62 = (lex.Digit | temp61);
243        temp63 = (temp16 & temp2);
244        temp64 = (temp63 &~ basis_bits.bit_4);
245        temp65 = (temp64 & temp60);
246        lex.Hex = (temp62 | temp65);
247        lex.error = x00_x1F &~ lex.WS
248       
249        ### Validate_utf8(basis_bits, u8):
250        u8.unibyte = (~basis_bits.bit_0);
251        u8.suffix = 0
252        u8.error = 0
253        u8.FFFE_FFFF = 0
254        u8anyscope = 0 #local
255        if basis_bits.bit_0:
256                u8.prefix = (basis_bits.bit_0 & basis_bits.bit_1);
257                u8.prefix2 = (u8.prefix &~ basis_bits.bit_2);
258                u8.prefix3 = (u8.prefix & temp2);
259                u8.prefix4 = (u8.prefix & temp7);
260                u8.suffix = (basis_bits.bit_0 &~ basis_bits.bit_1);
261                temp66 = (u8.prefix &~ temp49);
262                temp67 = (temp21 | basis_bits.bit_6);
263                temp68 = (temp66 &~ temp67);
264                temp69 = (basis_bits.bit_5 & temp13);
265                temp70 = (basis_bits.bit_4 | temp69);
266                temp71 = (u8.prefix4 & temp70);
267                u8.badprefix = (temp68 | temp71);
268                u8.error = u8.badprefix
269                u8.scope22 = bitutil.Advance(u8.prefix2)
270                u8anyscope = u8.scope22
271                if u8.prefix3 | u8.prefix4:
272                        xE0 = (u8.prefix3 &~ temp36);
273                        xED = (u8.prefix3 & temp20);
274                        xF0 = (u8.prefix4 &~ temp36);
275                        temp72 = (temp4 &~ temp13);
276                        xF4 = (u8.prefix4 & temp72);
277                        u8.xA0_xBF = (u8.suffix & basis_bits.bit_2);
278                        u8.x80_x9F = (u8.suffix &~ basis_bits.bit_2);
279                        u8.x90_xBF = (u8.suffix & temp49);
280                        u8.x80_x8F = (u8.suffix &~ temp49);
281                        xEF = (u8.prefix3 & temp23);
282                        temp73 = (u8.suffix & temp7);
283                        u8.xBF = (temp73 & temp23);
284                        u8.xBE = (temp73 & temp15);
285                        u8.xE0_scope = bitutil.Advance(xE0);
286                        u8.xED_scope = bitutil.Advance(xED);
287                        u8.xF0_scope = bitutil.Advance(xF0);
288                        u8.xF4_scope = bitutil.Advance(xF4);
289                        u8.xEF_scope = bitutil.Advance(xEF);
290                        u8.scope32 = bitutil.Advance(u8.prefix3)
291                        u8.scope33 = bitutil.Advance(u8.scope32)
292                        u8.scope42 = bitutil.Advance(u8.prefix4)
293                        u8.scope43 = bitutil.Advance(u8.scope42)
294                        u8.scope44 = bitutil.Advance(u8.scope43)
295
296                        u8lastscope = u8.scope22 | u8.scope33 | u8.scope44
297                        u8anyscope = u8lastscope | u8.scope32 | u8.scope42 | u8.scope43
298               
299                        u8error1 = u8.xE0_scope & u8.x80_x9F
300                        u8error2 = u8.xED_scope & u8.xA0_xBF
301                        u8error3 = u8.xF0_scope & u8.x80_x8F
302                        u8error4 = u8.xF4_scope & u8.x90_xBF
303       
304                        u8.error |= u8error1 | u8error2 | u8error3 | u8error4
305
306                        EF_BF_pending = bitutil.Advance(u8.xEF_scope & u8.xBF)
307
308                        u8.FFFE_FFFF = (EF_BF_pending & (u8.xBE | u8.xBF))
309        u8mismatch = u8anyscope ^ u8.suffix
310        u8.error |= u8mismatch
311       
312def Add_scope_streams(lex, scope1):
313        #scope1.LAngle = bitutil.Advance(lex.LAngle)
314        #scope1.Hyphen = bitutil.Advance(lex.Hyphen)
315        #scope1.QMark = bitutil.Advance(lex.QMark)
316        v = lex.LAngle | lex.Hyphen
317        w = lex.Hyphen | lex.QMark
318        v1 = bitutil.Advance(v)
319        w1 = bitutil.Advance(w)
320        scope1.LAngle = v1 &~ w1
321        scope1.Hyphen = v1 & w1
322        scope1.QMark = w1 &~ v1
323
324def Parse_CtCDPI(ctCDPI_Callouts, lex, scope1, check_streams):
325        ctCDPI_Callouts.CD_end = 0
326        ctCDPI_Callouts.Ct_starts = 0
327        ctCDPI_Callouts.Ct_ends = 0
328        ctCDPI_Callouts.CD_starts = 0
329        ctCDPI_Callouts.CD_ends = 0
330        ctCDPI_Callouts.PI_starts = 0
331        ctCDPI_Callouts.PI_name_starts = 0
332        ctCDPI_Callouts.PI_name_ends = 0
333        ctCDPI_Callouts.PI_ends = 0
334        ctCDPI_Callouts.CtCDPI_mask = 0
335        ctCDPI_Callouts.error = 0
336        CtCDPI_starts = 0
337        Ct_errors = 0
338
339        if lex.RBracket:
340                scope1.RBracket = bitutil.Advance(lex.RBracket)
341                ctCDPI_Callouts.CD_end = bitutil.Advance(scope1.RBracket & lex.RBracket) & lex.RAngle
342        PI_start = scope1.LAngle & lex.QMark
343        CtCD_start = scope1.LAngle & lex.Exclam
344        CtCDPI_start = PI_start | CtCD_start
345
346        DoubleHyphen = scope1.Hyphen & lex.Hyphen
347        PI_end = scope1.QMark & lex.RAngle
348
349
350        #
351        # Initiate the scan
352        CtCDPI_Cursor = bitutil.ScanToFirst(CtCDPI_start)
353        while CtCDPI_Cursor:
354                CtCDPI_starts |= CtCDPI_Cursor
355                PI_Cursor = CtCDPI_Cursor & PI_start
356                CD_Ct_Cursor = bitutil.Advance(CtCDPI_Cursor & ~PI_Cursor)
357                CD_Cursor = CD_Ct_Cursor & lex.LBracket
358                Ct_Cursor = CD_Ct_Cursor & lex.Hyphen
359                ctCDPI_Callouts.PI_starts |= PI_Cursor
360                ctCDPI_Callouts.CD_starts |= CD_Cursor
361                ctCDPI_Callouts.Ct_starts |= Ct_Cursor
362                Ct_Cursor = bitutil.Advance(Ct_Cursor) 
363                Ct_errors |= Ct_Cursor & ~ lex.Hyphen
364                # Advance twice past <!--, so that we don't treat <!---
365                # as being a terminated comment.
366                Ct_Cursor = bitutil.Advance(bitutil.Advance(Ct_Cursor))
367                PI_Cursor = bitutil.Advance(PI_Cursor)
368                ctCDPI_Callouts.PI_name_starts |= PI_Cursor
369                PI_name_end = bitutil.ScanThru(PI_Cursor, lex.NameScan)
370                ctCDPI_Callouts.PI_name_ends |= PI_name_end
371                PI_Cursor = bitutil.ScanTo(PI_name_end, PI_end)
372                CD_Cursor = bitutil.ScanTo(CD_Cursor, ctCDPI_Callouts.CD_end)
373                Ct_Cursor = bitutil.Advance(bitutil.ScanTo(Ct_Cursor, DoubleHyphen))
374                ctCDPI_Callouts.PI_ends |= PI_Cursor
375                ctCDPI_Callouts.CD_ends |= CD_Cursor
376                ctCDPI_Callouts.Ct_ends |= Ct_Cursor
377                CtCDPI_Cursor = PI_Cursor | CD_Cursor | Ct_Cursor
378                CtCDPI_Cursor = bitutil.ScanTo(CtCDPI_Cursor, CtCDPI_start)
379       
380                ctCDPI_Callouts.CtCDPI_mask = bitutil.Advance(ctCDPI_Callouts.CD_ends | ctCDPI_Callouts.Ct_ends | ctCDPI_Callouts.PI_ends) - CtCDPI_starts             
381                #ctCDPI_Callouts.error = Ct_ends & ~lex.RAngle | Ct_starts & ~ lex.Hyphen
382                ctCDPI_Callouts.error = Ct_errors | ctCDPI_Callouts.Ct_ends & ~lex.RAngle
383                ctCDPI_Callouts.error |= bitutil.Advance(ctCDPI_Callouts.PI_name_ends & ~ lex.WS) & ~ PI_end
384                ctCDPI_Callouts.error |= ctCDPI_Callouts.PI_name_starts & ctCDPI_Callouts.PI_name_ends
385                # If any of the Comment, CDATA or PI markups are unterminated, it is an error.
386                ctCDPI_Callouts.error |= ctCDPI_Callouts.CtCDPI_mask &~ EOF_mask
387               
388        check_streams.misc_mask = (lex.WS | lex.LAngle | (bitutil.Advance(ctCDPI_Callouts.Ct_ends | ctCDPI_Callouts.PI_ends) - (ctCDPI_Callouts.Ct_starts | ctCDPI_Callouts.PI_starts)) | CtCDPI_starts) & EOF_mask
389
390def Parse_tags(lex, scope1, ctCDPI_Callouts, tag_Callouts):
391
392       
393        # Delimiters for scans.
394        DQuoteDelim = lex.DQuote | lex.LAngle
395        SQuoteDelim = lex.SQuote | lex.LAngle
396        AttListDelim = lex.Slash | lex.RAngle
397       
398        # Start the parallel parsing by inspecting the character
399        # after the opening "<" of a tag.
400        tag_Callouts.LAngleFollow = scope1.LAngle &~ ctCDPI_Callouts.CtCDPI_mask
401        tag_Callouts.ElemName_starts = tag_Callouts.LAngleFollow & ~lex.Slash
402        tag_Callouts.EndTag_marks = tag_Callouts.LAngleFollow & lex.Slash
403       
404        # Start Tag/Empty Element Tag Parsing
405
406        # Advance all cursors by scanning through the tag name.
407        tag_Callouts.ElemName_ends = bitutil.ScanThru(tag_Callouts.ElemName_starts, lex.NameScan)
408        # Must have at least one name character for a legal start tag.
409        # Mark any occurrences of null names as errors.
410        ParseError = tag_Callouts.ElemName_starts & tag_Callouts.ElemName_ends
411       
412        # Initialize the accumulators for attribute name and value positions.
413        tag_Callouts.AttName_starts = 0 
414        tag_Callouts.AttName_ends = 0
415        EqToCheck = 0
416        tag_Callouts.AttVal_starts = 0
417        AttValEnds = 0
418        tag_Callouts.AttVal_ends = 0
419
420        # After the element name, there may or may not be an attlist.
421        AfterWS = bitutil.ScanThru(tag_Callouts.ElemName_ends, lex.WS)
422        AttListEnd = AfterWS & AttListDelim
423        AttNameStart = AfterWS & ~AttListDelim
424        # At least one WS character is required between ElemNames and AttNames.
425        ParseError |= tag_Callouts.ElemName_ends & AttNameStart
426
427        #
428        # The following loop iterates through attributes within a start tag.
429        # Because all start tags are processed in parallel, the number of
430        # iterations is the maximum number of attributes found in any one
431        # start tag, plus one.
432        while AttNameStart:
433                tag_Callouts.AttName_starts |= AttNameStart
434                AttNameFollow = bitutil.ScanThru(AttNameStart, lex.NameScan)
435                tag_Callouts.AttName_ends |= AttNameFollow
436                # Scan through WS to the expected '=' delimiter.
437                EqExpected = bitutil.ScanThru(AttNameFollow, lex.WS)
438                EqToCheck |= EqExpected
439                AttValPos = bitutil.ScanThru(EqExpected, EqExpected | lex.WS)
440                tag_Callouts.AttVal_starts |= AttValPos
441                DQuoteAttVal = AttValPos & lex.DQuote
442                SQuoteAttVal = AttValPos & lex.SQuote
443                DQuoteAttEnd = bitutil.ScanTo(DQuoteAttVal, DQuoteDelim &~ DQuoteAttVal)
444                SQuoteAttEnd = bitutil.ScanTo(SQuoteAttVal, SQuoteDelim &~ SQuoteAttVal)
445                AttValEnd = DQuoteAttEnd | SQuoteAttEnd
446                AttValEnds |= AttValEnd
447                AttValFollow = bitutil.Advance(AttValEnd)
448                tag_Callouts.AttVal_ends |= AttValFollow
449                AfterWS = bitutil.ScanThru(AttValFollow, lex.WS)
450                AttListEnd |= AfterWS & AttListDelim
451                AttNameStart = AfterWS & ~AttListDelim
452
453        # No more attribute values to process when AttNameStart == 0.
454        STagEnds = AttListEnd & lex.RAngle
455        # Mark any "/" characters found as the ends of empty element tags.
456        tag_Callouts.EmptyTag_marks = bitutil.Advance(AttListEnd & lex.Slash)
457       
458        # Check for errors.
459        ParseError |= tag_Callouts.AttVal_ends & tag_Callouts.AttName_starts # No intervening WS.
460        ParseError |= tag_Callouts.AttName_starts & tag_Callouts.AttName_ends # Null AttName
461        ParseError |= EqToCheck & ~lex.Equals # = not found where expected.
462        ParseError |= tag_Callouts.AttVal_starts & ~ (lex.DQuote | lex.SQuote)
463        ParseError |= AttValEnds & ~ (lex.DQuote | lex.SQuote)
464        ParseError |= tag_Callouts.EmptyTag_marks & ~lex.RAngle
465
466        # End Tag Parsing
467        EndTagEnds = bitutil.ScanThru(bitutil.ScanThru(tag_Callouts.EndTag_marks, tag_Callouts.EndTag_marks | lex.NameScan), lex.WS)
468        ParseError |= EndTagEnds & ~lex.RAngle
469        tag_Callouts.error = ParseError
470               
471        # Attribute value spans
472        tag_Callouts.AttVal_spans = tag_Callouts.AttVal_ends - tag_Callouts.AttVal_starts
473                       
474def Parse_refs(lex, scope1, ctCDPI_Callouts, ref_Callouts):
475        ref_Callouts.GenRef_starts = 0
476        ref_Callouts.GenRef_ends = 0
477        ref_Callouts.DecRef_starts = 0
478        ref_Callouts.DecRef_ends = 0
479        ref_Callouts.HexRef_starts = 0
480        ref_Callouts.HexRef_ends = 0
481        ref_Callouts.error = 0
482
483        Ref1 = lex.RefStart &~ ctCDPI_Callouts.CtCDPI_mask
484        # All remaining "&" must be reference start characters; parse them.
485        if Ref1:
486                scope1.RefStart = bitutil.Advance(Ref1)
487                NumRef2 = scope1.RefStart & lex.Hash
488                ref_Callouts.GenRef_starts = scope1.RefStart &~ lex.Hash
489                NumRef3 = bitutil.Advance(NumRef2)
490                HexRef3 = NumRef3 & lex.x
491                ref_Callouts.DecRef_starts = NumRef3 &~ lex.x
492                ref_Callouts.HexRef_starts = bitutil.Advance(HexRef3) 
493                ref_Callouts.GenRef_ends = bitutil.ScanThru(ref_Callouts.GenRef_starts, lex.NameScan)
494                ref_Callouts.DecRef_ends = bitutil.ScanThru(ref_Callouts.DecRef_starts, lex.Digit)
495                ref_Callouts.HexRef_ends = bitutil.ScanThru(ref_Callouts.HexRef_starts, lex.Hex)
496                # Error checks
497                # At least one digit required for DecRef, one hex digit for HexRef.
498                ref_error1 = ref_Callouts.DecRef_starts &~ lex.Digit
499                ref_error2 = ref_Callouts.HexRef_starts &~ lex.Hex
500                # Semicolon terminator required (also covers unterminated at EOF).
501                ref_ends = ref_Callouts.GenRef_ends | ref_Callouts.DecRef_ends | ref_Callouts.HexRef_ends
502                ref_error3 = ref_ends &~ lex.Semicolon
503                ref_Callouts.error = ref_error1 | ref_error2 | ref_error3
504
505def Validate_xml_names(ctCDPI_Callouts, ref_Callouts, tag_Callouts, lex, u8, xml_names, check_streams):
506        PI_names = ctCDPI_Callouts.PI_name_ends - ctCDPI_Callouts.PI_name_starts
507        GenRefs = ref_Callouts.GenRef_ends - ref_Callouts.GenRef_starts
508        ElemNames = tag_Callouts.ElemName_ends - tag_Callouts.ElemName_starts
509        AttNames = tag_Callouts.AttName_ends - tag_Callouts.AttName_starts
510        qname_stream =  ElemNames | AttNames
511        ncname_stream = PI_names | GenRefs
512        name_stream = qname_stream | ncname_stream
513        name_start = name_stream &~ bitutil.Advance(name_stream)
514        name_cursor = name_stream & ~bitutil.Advance(name_stream)
515        void_prefix_err = name_cursor & lex.Colon
516        namespace_sep = bitutil.ScanThru(name_cursor, lex.NameScan &~ lex.Colon) & lex.Colon
517        local_part_start = bitutil.Advance(namespace_sep)
518        local_part_err = local_part_start &~ lex.NameScan
519        colon2_err = bitutil.ScanThru(local_part_start, lex.NameScan &~ lex.Colon) & lex.Colon
520        ncname_err = ncname_stream & lex.Colon
521        xml_names.namespace_error = void_prefix_err | local_part_err | colon2_err | ncname_err
522                       
523        check_streams.non_ascii_name_starts = name_start &~lex.ASCII_name_start
524        check_streams.non_ascii_names = (name_stream &~ name_start) & ~lex.ASCII_name_char & ~u8.suffix
525   
526def Do_check_streams(ctCDPI_Callouts, tag_Callouts, lex, u8, scope1, ref_Callouts, xml_names, check_streams):
527    # Ensure that no occurrence of ]]> occurs outside of markup.
528    CD_end_error = ctCDPI_Callouts.CD_end & ~(ctCDPI_Callouts.CtCDPI_mask | tag_Callouts.AttVal_spans)
529           
530    # Consolidate and check for errors
531    check_streams.error_mask = lex.error & EOF_mask | u8.error | u8.FFFE_FFFF | ctCDPI_Callouts.error | tag_Callouts.error | CD_end_error | ref_Callouts.error | xml_names.namespace_error
532
533    check_streams.tag_marks = tag_Callouts.EmptyTag_marks | tag_Callouts.LAngleFollow | tag_Callouts.AttName_starts
534    check_streams.name_follows = tag_Callouts.ElemName_ends | tag_Callouts.AttName_ends
535    check_streams.att_refs = tag_Callouts.AttVal_spans & scope1.RefStart
536
537def Form_Length_Group_Bitstreams(tag_Callouts):
538
539    starts = tag_Callouts.ElemName_starts
540    remaining_ends = tag_Callouts.ElemName_ends
541
542    # Group symbols of length 1
543    starts_1 = bitutil.Advance(starts)
544    tag_Callouts.ElemName_ends_1 = starts_1 & remaining_ends
545    remaining_ends = remaining_ends & ~tag_Callouts.ElemName_ends_1
546
547    # Group symbols of length 2
548    starts_2 = bitutil.Advance(starts_1)
549    tag_Callouts.ElemName_ends_2 = starts_2 & remaining_ends
550    remaining_ends = remaining_ends & ~tag_Callouts.ElemName_ends_2
551
552    # Group symbols of length 3_to_4
553    temp = starts_1 | starts_2
554    temp32 = bitutil.Advance32(temp)
555    #starts_3_to_4 = bitutil.Advance(bitutil.Advance(temp))
556    starts_3_to_4 = interpose32 (temp, temp32, 2)
557    tag_Callouts.ElemName_ends_3_to_4 = starts_3_to_4 & remaining_ends
558    remaining_ends = remaining_ends & ~tag_Callouts.ElemName_ends_3_to_4
559
560    # Group symbols of length 5_to_8
561    temp = temp | starts_3_to_4
562    temp32 = bitutil.Advance32(temp)
563    #starts_5_to_8 = bitutil.Advance(bitutil.Advance(bitutil.Advance(bitutil.Advance(temp))))
564    starts_5_to_8 = interpose32 (temp, temp32, 4)
565    tag_Callouts.ElemName_ends_5_to_8 = starts_5_to_8 & remaining_ends
566    remaining_ends = remaining_ends & ~tag_Callouts.ElemName_ends_5_to_8
567
568    # Group symbols of length 9_to_16
569    temp = temp | starts_5_to_8
570    temp32 = bitutil.Advance32 (temp)
571    #starts_9_to_16 = bitutil.Advance(bitutil.Advance(bitutil.Advance(bitutil.Advance(bitutil.Advance(bitutil.Advance(bitutil.Advance(bitutil.Advance(temp))))))))
572    starts_9_to_16 = interpose32 (temp, temp32, 8)
573    tag_Callouts.ElemName_ends_9_to_16 = starts_9_to_16 & remaining_ends
574    remaining_ends = remaining_ends & ~tag_Callouts.ElemName_ends_9_to_16
575
576    # Group symbols of length 17 and longer
577    tag_Callouts.ElemName_ends_17_and_longer = remaining_ends
578
579
580def Compute_Hash_Value_Bitstream(hash_data, basis_bits):
581    hash_data.Hash_value = basis_bits.bit_2 ^ basis_bits.bit_4 ^ basis_bits.bit_6
582    #hash_data.Hash_value = basis_bits.bit_3 ^ basis_bits.bit_5 ^ basis_bits.bit_7
583
584#def main(basis_bits, lex, u8, scope1, ctCDPI_Callouts, masks, check_streams, tag_Callouts, ref_Callouts, xml_names):
585def Main(basis_bits, lex, u8, xml_char, scope1, ctCDPI_Callouts, ref_Callouts, tag_Callouts, masks, xml_names, check_streams, hash_data):
586       
587        # Classify bytes for UTF-8 processing, whitespace and control
588        # processing and XML lexical analysis.
589        # Classify_bytes(basis_bits, lex)
590
591        # Validate UTF-8 multibyte sequences and determine the UTF-8 scope streams
592        # Validate_utf8(basis_bits, u8)
593                               
594        Classify_bytes_Validate_utf8(basis_bits, lex, u8)
595
596        Add_scope_streams(lex, scope1)
597   
598        # Parse all comments, CDATA sections and processing instructions.
599        Parse_CtCDPI(ctCDPI_Callouts, lex, scope1, check_streams)
600               
601        # All remaining '<' must be tag start characters; parse tags.
602        Parse_tags(lex, scope1, ctCDPI_Callouts, tag_Callouts) 
603
604        # All remaining '&' must be reference start characters; parse them.
605        Parse_refs(lex, scope1, ctCDPI_Callouts, ref_Callouts)
606       
607        # Validate XML namespaces and generate bit streams to post validate non-ascii range XML names
608        Validate_xml_names(ctCDPI_Callouts, ref_Callouts, tag_Callouts, lex, u8, xml_names, check_streams)
609   
610   
611        Do_check_streams(ctCDPI_Callouts, tag_Callouts, lex, u8, scope1, ref_Callouts, xml_names, check_streams)
612
613        # These methods are needed to do Paralel Bitstream Based Length Sorting
614        Form_Length_Group_Bitstreams(tag_Callouts)
615
616        Compute_Hash_Value_Bitstream(hash_data, basis_bits);
617               
Note: See TracBrowser for help on using the repository browser.