root/proto/parabix2/pablo_template.cpp @ 1761

Revision 1761, 12.9 KB (checked in by ksherdy, 19 months ago)

Revert s2p 128,256.
Line 
1#include <stdio.h>
2#include <stdlib.h>
3#include <errno.h>
4#include <sys/types.h>
5#include <sys/stat.h>
6
7#define LocalCarryDeclare(name, count)\
8CarryArray<count> name;\
9
10#include "../lib/bitblock.hpp"
11#include "../lib/carryQ.hpp"
12#include "../lib/bitstream_iterator.hpp"
13#include "../lib/bitblock_iterator.hpp"
14#include "../lib/s2p.hpp"
15
16#define SEGMENT_BLOCKS 12
17#define BUFFER_SIZE (BLOCK_SIZE * SEGMENT_BLOCKS)
18#define OVERLAP_BUFSIZE (sizeof(BitBlock))
19
20#include "xmldecl.h"
21#include "namechars.h"
22#include "../lib/perflib/perfsec.h"
23
24#include "TagMatcher.hpp"
25#include "LineColTracker.hpp"
26#include "ErrorUtil.h"
27#include "ErrorTracker.h"
28#include "XMLTestSuiteError.h"
29
30#ifdef BUFFER_PROFILING
31        BOM_Table * parser_timer;
32
33#elif CODE_CLOCKER
34        #define NUM_EVENTS 1
35        int Events[NUM_EVENTS] = {PAPI_TOT_CYC};
36        //int Events[NUM_EVENTS] = {PAPI_L2_DCM};
37        //int Events[NUM_EVENTS] = {PAPI_TOT_CYC, PAPI_BR_MSP};
38        int cal_size = 20;
39        CC * parser_timer = new CC(Events,NUM_EVENTS,cal_size);
40#else
41        void * parser_timer;
42#endif
43
44int block_base=0;
45int buffer_base=0;
46char * source;
47
48LineColTracker tracker;
49TagMatcher matcher;
50ErrorTracker error_tracker;
51BitBlock EOF_mask = simd<1>::constant<1>();
52
53static inline int NameStrt_check(int pos);
54static inline int Name_check(int pos);
55static inline int PIName_check(int pos);
56static inline int CD_check(int pos);
57static inline int GenRef_check(int pos);
58static inline int HexRef_check(int pos);
59static inline int DecRef_check(int pos);
60static inline int AttRef_check(int pos);
61
62@global
63
64static inline void s2p_do_block(BytePack U8[], Basis_bits & basis_bits);
65static inline void s2p_do_final_block(BytePack U8[], Basis_bits & basis_bits, BitBlock EOF_mask);
66static inline void postprocess_do_block(Lex & lex, CtCDPI_Callouts & ctCDPI_Callouts, Ref_Callouts & ref_Callouts, Check_streams & check_streams, int chars_avail);
67
68void do_process(FILE *infile, FILE *outfile);
69
70static inline void validate_block(BitBlockForwardIterator & start, int block_base, int is_valid(int));
71static inline void validate_block(BitBlockForwardIterator & start, int block_base, int is_valid(int,int));
72
73int main(int argc, char * argv[]) {
74
75        char * infilename, * outfilename;
76        FILE *infile, *outfile;
77        struct stat fileinfo;
78
79        if (argc < 2) {
80                printf("Usage: %s <filename> [<outputfile>]\n", argv[0]);
81                exit(-1);
82        }
83
84        infilename = argv[1];
85        stat(infilename, &fileinfo);
86        infile = fopen(infilename, "rb");
87        if (!infile) {
88                fprintf(stderr, "Error: cannot open %s for input.\n", infilename);
89                exit(-1);
90        }
91
92        if (argc < 3) outfile = stdout;
93        else {
94                outfilename = argv[2];
95                outfile = fopen(outfilename, "wb");
96                if (!outfile) {
97                        fprintf(stderr, "Error: cannot open %s for writing.\n", outfilename);
98                        exit(-1);
99                }
100        }
101
102//      PERF_SEC_BIND(1);
103
104        PERF_SEC_INIT(parser_timer);
105
106        do_process(infile, outfile);
107
108        PERF_SEC_DUMP(parser_timer);
109
110        PERF_SEC_DESTROY(parser_timer);
111
112        fclose(infile);
113        fclose(outfile);
114
115        return(0);
116}
117
118/* s2p Definitions */
119static inline void s2p_do_block(BytePack U8[], Basis_bits & basis_bits) {
120  s2p(U8[0], U8[1], U8[2], U8[3], U8[4], U8[5], U8[6], U8[7],
121        basis_bits.bit_0, basis_bits.bit_1, basis_bits.bit_2, basis_bits.bit_3, basis_bits.bit_4, basis_bits.bit_5, basis_bits.bit_6, basis_bits.bit_7);
122}
123
124static inline void s2p_do_final_block(BytePack U8[], Basis_bits & basis_bits, BitBlock EOF_mask) {
125  s2p_do_block(U8, basis_bits);
126  basis_bits.bit_0 = simd_and(basis_bits.bit_0, EOF_mask);
127  basis_bits.bit_1 = simd_and(basis_bits.bit_1, EOF_mask);
128  basis_bits.bit_2 = simd_and(basis_bits.bit_2, EOF_mask);
129  basis_bits.bit_3 = simd_and(basis_bits.bit_3, EOF_mask);
130  basis_bits.bit_4 = simd_and(basis_bits.bit_4, EOF_mask);
131  basis_bits.bit_5 = simd_and(basis_bits.bit_5, EOF_mask);
132  basis_bits.bit_6 = simd_and(basis_bits.bit_6, EOF_mask);
133  basis_bits.bit_7 = simd_and(basis_bits.bit_7, EOF_mask);
134}
135
136
137static inline int NameStrt_check(int pos) {
138        if(XML_10_UTF8_NameStrt_bytes((unsigned char*)&source[pos]) == 0){
139              return XMLTestSuiteError::NAME_START;
140        }
141        return 0;
142}
143
144static inline int Name_check(int pos) {
145        if(XML_10_UTF8_NameChar_bytes((unsigned char*)&source[pos]) == 0){
146                  return XMLTestSuiteError::NAME;
147        }
148        return 0;
149}
150
151static inline int PIName_check(int pos, int file_pos) {
152        if (at_XxMmLll<ASCII>((unsigned char*)&source[pos]) && (source[pos+3]=='?' || source[pos+3]<= ' ')) {
153              // "<?xml" legal at start of file.
154              if (!((file_pos == 2) && at_XmlDecl_start<ASCII>((unsigned char*)&source[0]))) {
155                  return XMLTestSuiteError::XMLPINAME;
156              }
157        }
158        return 0;
159}
160
161static inline int CD_check(int pos) {
162        if (!at_CDATA1<ASCII>((unsigned char*)&source[pos])){
163                  return XMLTestSuiteError::CDATA;
164        }
165        return 0;
166}
167
168static inline int GenRef_check(int pos) {
169        unsigned char* s = (unsigned char*)&source[pos];
170        if (!(at_Ref_gt<ASCII>(s)||at_Ref_lt<ASCII>(s)||at_Ref_amp<ASCII>(s)||at_Ref_quot<ASCII>(s)||at_Ref_apos<ASCII>(s))){
171              return XMLTestSuiteError::UNDEFREF;
172        }
173        return 0;
174}
175
176static inline int HexRef_check(int pos) {
177        unsigned char* s = (unsigned char*)&source[pos];
178        int ch_val = 0;
179        while(at_HexDigit<ASCII>(s)){
180          ch_val = HexVal<ASCII>(s[0]) + (ch_val<<4);
181          if (ch_val> 0x10FFFF ){
182                return XMLTestSuiteError::CHARREF;
183          }
184          s++;
185        }
186        if ((ch_val == 0x0) || ((ch_val | 0x7FF) == 0xDFFF)|| ((ch_val | 0x1) == 0xFFFF)){
187          return XMLTestSuiteError::CHARREF;
188        }
189        else if (((ch_val < 0x20) && (ch_val != 0x9) && (ch_val != 0xD) && (ch_val != 0xA))){
190          return XMLTestSuiteError::XML10CHARREF;
191        }
192        return 0;
193}
194
195static inline int DecRef_check(int pos) {
196        unsigned char* s = (unsigned char*)&source[pos];
197        int ch_val = 0;
198        while(at_HexDigit<ASCII>(s)){
199          ch_val = DigitVal<ASCII>(s[0]) + ch_val*10;
200          if (ch_val> 0x10FFFF ){
201                        return XMLTestSuiteError::CHARREF;
202          }
203          s++;
204        }
205        if ((ch_val == 0x0) || ((ch_val | 0x7FF) == 0xDFFF)|| ((ch_val | 0x1) == 0xFFFF)){
206                  return XMLTestSuiteError::CHARREF;
207        }
208        else if (((ch_val < 0x20) && (ch_val != 0x9) && (ch_val != 0xD) && (ch_val != 0xA))){
209                  return XMLTestSuiteError::XML10CHARREF;
210        }
211        return 0;
212}
213
214static inline int AttRef_check(int pos) {
215        unsigned char* s = (unsigned char*)&source[pos];
216        int ch_val = 0;
217        if(s[0]=='#'){
218          s++;
219          if(s[0]=='x' || s[0]=='X'){
220            s++;
221            while(at_HexDigit<ASCII>(s)){
222              ch_val = HexVal<ASCII>(s[0]) + (ch_val<<4);
223              s++;
224            }
225          }
226          else{
227            while(at_HexDigit<ASCII>(s)){
228              ch_val = DigitVal<ASCII>(s[0]) + ch_val*10;
229              s++;
230            }
231          }
232          if (ch_val==60){
233            return XMLTestSuiteError::ATTREF;
234          }
235        }
236        else if(at_Ref_lt<ASCII>(s)){
237          return XMLTestSuiteError::ATTREF;
238        }
239        return 0;
240}
241
242static inline void validate_block(BitBlockForwardIterator & start, int block_base, int is_valid(int)) {
243
244        int pos, block_pos;
245        BitBlockForwardIterator end;
246        while(start != end) {
247
248                block_pos = block_base + *start;
249                int rv = is_valid(block_pos);
250
251                if (rv) {
252                        int error_line, error_column;
253                        tracker.get_Line_and_Column(block_pos, error_line, error_column);
254                        ReportError(XMLTestSuiteError::get_msg(rv), error_line, error_column);
255                        exit(-1);
256                }
257                start++;
258        }
259}
260
261static inline void validate_block(BitBlockForwardIterator & start, int block_base, int buffer_base, int is_valid(int,int)) {
262
263        int pos, block_pos, file_pos;
264        BitBlockForwardIterator end;
265        while(start != end) {
266
267                block_pos = block_base + *start;
268                file_pos = block_pos+buffer_base;
269
270
271                int rv = is_valid(block_pos, file_pos);
272
273                if (rv) {
274                        int error_line, error_column;
275                        tracker.get_Line_and_Column(block_pos, error_line, error_column);
276                        ReportError(XMLTestSuiteError::get_msg(rv), error_line, error_column);
277                        exit(-1);
278                }
279                start++;
280        }
281}
282
283static inline void postprocess_do_block(Lex & lex, CtCDPI_Callouts & ctCDPI_Callouts, Ref_Callouts & ref_Callouts, Check_streams & check_streams, int chars_avail){
284BitBlockForwardIterator iter;
285
286tracker.StoreNewlines(lex.LF);
287
288if (bitblock_has_bit(simd_or(check_streams.non_ascii_name_starts, check_streams.non_ascii_names))) {
289        iter.init(&check_streams.non_ascii_name_starts);
290        validate_block(iter, block_base, NameStrt_check);
291        iter.init(&check_streams.non_ascii_names);
292        validate_block(iter, block_base, Name_check);
293}
294if (bitblock_has_bit(ctCDPI_Callouts.PI_name_starts)){
295        iter.init(&(ctCDPI_Callouts.PI_name_starts));
296        validate_block(iter, block_base, buffer_base, PIName_check);
297}
298if (bitblock_has_bit(ctCDPI_Callouts.CD_starts)){
299        iter.init(&ctCDPI_Callouts.CD_starts);
300        validate_block(iter, block_base, CD_check);
301}
302if(bitblock_has_bit(ref_Callouts.GenRef_starts)){
303        iter.init(&ref_Callouts.GenRef_starts);
304        validate_block(iter, block_base, GenRef_check);
305}
306if(bitblock_has_bit(ref_Callouts.DecRef_starts)){
307        iter.init(&ref_Callouts.DecRef_starts);
308        validate_block(iter, block_base, DecRef_check);
309}
310if(bitblock_has_bit(ref_Callouts.HexRef_starts)){
311        iter.init(&ref_Callouts.HexRef_starts);
312        validate_block(iter, block_base, HexRef_check);
313}
314if(bitblock_has_bit(check_streams.att_refs)){
315        iter.init(&check_streams.att_refs);
316        validate_block(iter, block_base, AttRef_check);
317}
318
319if(error_tracker.Has_Noted_Error()){
320        int error_line, error_column;
321        tracker.get_Line_and_Column(error_tracker.Noted_Pos_In_Block(), error_line, error_column);
322        ReportError(error_tracker.Noted_Error_Msg(), error_line, error_column);
323        exit(-1);
324}
325
326matcher.store_streams(check_streams.tag_marks, check_streams.name_follows, check_streams.misc_mask, chars_avail);
327tracker.AdvanceBlock();
328
329}
330
331void do_process(FILE *infile, FILE *outfile) {
332
333@decl
334
335  int buf_pos = 0;
336  int block_pos = 0;
337  int chars_avail = 0;
338  int check_pos = 0;
339  int chars_read = 0;
340  BytePack buf[(BUFFER_SIZE+BLOCK_SIZE+OVERLAP_BUFSIZE*2)/sizeof(BitBlock)];
341
342  char * srcbuf = ((char *) buf) + OVERLAP_BUFSIZE;
343  buffer_base = buf_pos;
344  source = srcbuf;
345
346  chars_read = fread((void *)srcbuf, 1, BUFFER_SIZE + OVERLAP_BUFSIZE, infile);
347  chars_avail = chars_read;
348  if (chars_avail > BUFFER_SIZE) chars_avail = BUFFER_SIZE;
349
350  matcher.setSrc(srcbuf);
351
352  if(chars_read<4){
353    fprintf(stderr,"File is too short. Not well formed.\n");
354    exit(-1);
355  }
356
357  Entity_Info * e = new Entity_Info;
358  e->AnalyzeSignature((unsigned char *)srcbuf);
359
360  if (e->code_unit_base == ASCII) {
361
362    XML_Decl_Parser<ASCII> decl_parser((unsigned char *)srcbuf);
363
364    decl_parser.ReadXMLInfo(*e);
365
366    if (e->code_unit_size != SingleByte || (e->has_encoding_decl && (!at_UTF_8(e->encoding)))){
367        fprintf(stderr,"Sorry, this xmlwf demo only works for UTF-8.\n");
368        exit(-1);
369    }
370  }
371  else {
372    fprintf(stderr,"Sorry, this xmlwf demo does not process EBCDIC.\n");
373        exit(-1);
374  }
375
376  if (e->content_start != 0) {
377        memmove(&srcbuf[0], &srcbuf[e->content_start], chars_read - e->content_start);
378        buf_pos = e->content_start;
379        if (chars_avail == BUFFER_SIZE) {
380                chars_read = chars_read - e->content_start +
381                             fread(&srcbuf[chars_read-e->content_start], 1, e->content_start, infile);
382                chars_avail = chars_read;
383                if (chars_avail > BUFFER_SIZE) chars_avail = BUFFER_SIZE;
384        }
385        else {
386          chars_read -=e->content_start;
387          chars_avail -=e->content_start;
388        }
389  }
390
391@stream_stmts
392
393/* Full Buffers */
394
395    while (chars_avail == BUFFER_SIZE) {
396      PERF_SEC_START(parser_timer);
397      for (int blk = 0; blk < SEGMENT_BLOCKS; blk++) {
398          block_base = blk*BLOCK_SIZE;
399          s2p_do_block((BytePack *) &srcbuf[block_base], basis_bits);
400          @block_stmts
401          postprocess_do_block(lex, ctCDPI_Callouts, ref_Callouts, check_streams, chars_avail);
402      }
403      matcher.StreamScan(chars_avail);
404      matcher.Advance_buffer();
405      tracker.Advance_buffer();
406      PERF_SEC_END(parser_timer, chars_avail);
407
408      int bytes_left = chars_read - chars_avail;
409      memmove(srcbuf, &srcbuf[BUFFER_SIZE], bytes_left);
410      chars_read = fread(&srcbuf[bytes_left],1, BUFFER_SIZE + OVERLAP_BUFSIZE - bytes_left, infile) + bytes_left;
411      chars_avail = chars_read;
412      if (chars_avail > BUFFER_SIZE) chars_avail = BUFFER_SIZE;
413      buf_pos += chars_avail;
414      buffer_base = buf_pos;
415    }
416/* Final Partial Buffer */
417    PERF_SEC_START(parser_timer);
418
419    block_pos = 0;
420    int remaining = chars_avail;
421/* Full Blocks */
422    while (remaining >= BLOCK_SIZE) {
423          block_base = block_pos;
424          s2p_do_block((BytePack *) &srcbuf[block_pos], basis_bits);
425          @block_stmts
426          postprocess_do_block(lex, ctCDPI_Callouts, ref_Callouts, check_streams, chars_avail);
427          block_pos += BLOCK_SIZE;
428          remaining -= BLOCK_SIZE;
429    }
430    block_base = block_pos;
431    if (remaining > 0 || @any_carry) {
432          EOF_mask = bitblock::srl(simd<1>::constant<1>(), convert(BLOCK_SIZE-remaining));
433          s2p_do_final_block((BytePack *) &srcbuf[block_pos], basis_bits, EOF_mask);
434          @final_block_stmts
435          postprocess_do_block(lex, ctCDPI_Callouts, ref_Callouts, check_streams, chars_avail);
436    }
437    buf_pos += chars_avail;
438    buffer_base = buf_pos;
439
440    matcher.StreamScan(chars_avail);
441    matcher.Advance_buffer();
442    tracker.Advance_buffer();
443
444
445    PERF_SEC_END(parser_timer, chars_avail);
446    if (matcher.depth != 0) {
447      fprintf(stderr, "tag matching error (depth %i) at position %i\n", matcher.depth, buffer_base);
448      exit(-1);
449    }
450}
451
Note: See TracBrowser for help on using the browser.