1 | //
|
---|
2 | // Cforall Version 1.0.0 Copyright (C) 2016 University of Waterloo
|
---|
3 | //
|
---|
4 | // The contents of this file are covered under the licence agreement in the
|
---|
5 | // file "LICENCE" distributed with Cforall.
|
---|
6 | //
|
---|
7 | // string_res -- variable-length, mutable run of text, with resource semantics
|
---|
8 | //
|
---|
9 | // Author : Michael L. Brooks
|
---|
10 | // Created On : Fri Sep 03 11:00:00 2021
|
---|
11 | // Last Modified By : Michael L. Brooks
|
---|
12 | // Last Modified On : Fri Sep 03 11:00:00 2021
|
---|
13 | // Update Count : 1
|
---|
14 | //
|
---|
15 |
|
---|
16 | #include "string_res.hfa"
|
---|
17 | #include <stdlib.hfa> // e.g. malloc
|
---|
18 | #include <string.h> // e.g. strlen
|
---|
19 |
|
---|
20 | //######################### VbyteHeap "header" #########################
|
---|
21 |
|
---|
22 |
|
---|
23 |
|
---|
24 |
|
---|
25 |
|
---|
26 |
|
---|
27 |
|
---|
28 |
|
---|
29 | // DON'T COMMIT:
|
---|
30 | // #define VbyteDebug
|
---|
31 |
|
---|
32 |
|
---|
33 |
|
---|
34 |
|
---|
35 |
|
---|
36 | #ifdef VbyteDebug
|
---|
37 | HandleNode *HeaderPtr;
|
---|
38 | #endif // VbyteDebug
|
---|
39 |
|
---|
40 | struct VbyteHeap {
|
---|
41 |
|
---|
42 | int NoOfCompactions; // number of compactions of the byte area
|
---|
43 | int NoOfExtensions; // number of extensions in the size of the byte area
|
---|
44 | int NoOfReductions; // number of reductions in the size of the byte area
|
---|
45 |
|
---|
46 | int InitSize; // initial number of bytes in the byte-string area
|
---|
47 | int CurrSize; // current number of bytes in the byte-string area
|
---|
48 | char *StartVbyte; // pointer to the `st byte of the start of the byte-string area
|
---|
49 | char *EndVbyte; // pointer to the next byte after the end of the currently used portion of byte-string area
|
---|
50 | void *ExtVbyte; // pointer to the next byte after the end of the byte-string area
|
---|
51 |
|
---|
52 | HandleNode Header; // header node for handle list
|
---|
53 | }; // VbyteHeap
|
---|
54 |
|
---|
55 |
|
---|
56 | static inline void compaction( VbyteHeap & ); // compaction of the byte area
|
---|
57 | static inline void garbage( VbyteHeap & ); // garbage collect the byte area
|
---|
58 | static inline void extend( VbyteHeap &, int ); // extend the size of the byte area
|
---|
59 | static inline void reduce( VbyteHeap &, int ); // reduce the size of the byte area
|
---|
60 |
|
---|
61 | static inline void ?{}( VbyteHeap &, int = 1000 );
|
---|
62 | static inline void ^?{}( VbyteHeap & );
|
---|
63 | static inline void ByteCopy( VbyteHeap &, char *, int, int, char *, int, int ); // copy a block of bytes from one location in the heap to another
|
---|
64 | static inline int ByteCmp( VbyteHeap &, char *, int, int, char *, int, int ); // compare 2 blocks of bytes
|
---|
65 | static inline char *VbyteAlloc( VbyteHeap &, int ); // allocate a block bytes in the heap
|
---|
66 |
|
---|
67 |
|
---|
68 | static inline void AddThisAfter( HandleNode &, HandleNode & );
|
---|
69 | static inline void DeleteNode( HandleNode & );
|
---|
70 | static inline void MoveThisAfter( HandleNode &, const HandleNode & ); // move current handle after parameter handle
|
---|
71 |
|
---|
72 |
|
---|
73 | // Allocate the storage for the variable sized area and intialize the heap variables.
|
---|
74 |
|
---|
75 | static inline void ?{}( VbyteHeap & this, int Size ) with(this) {
|
---|
76 | #ifdef VbyteDebug
|
---|
77 | serr | "enter:VbyteHeap::VbyteHeap, this:" | &this | " Size:" | Size;
|
---|
78 | #endif // VbyteDebug
|
---|
79 | NoOfCompactions = NoOfExtensions = NoOfReductions = 0;
|
---|
80 | InitSize = CurrSize = Size;
|
---|
81 | StartVbyte = EndVbyte = alloc(CurrSize);
|
---|
82 | ExtVbyte = (void *)( StartVbyte + CurrSize );
|
---|
83 | Header.flink = Header.blink = &Header;
|
---|
84 | #ifdef VbyteDebug
|
---|
85 | HeaderPtr = &Header;
|
---|
86 | serr | "exit:VbyteHeap::VbyteHeap, this:" | &this;
|
---|
87 | #endif // VbyteDebug
|
---|
88 | } // VbyteHeap
|
---|
89 |
|
---|
90 |
|
---|
91 | // Release the dynamically allocated storage for the byte area.
|
---|
92 |
|
---|
93 | static inline void ^?{}( VbyteHeap & this ) with(this) {
|
---|
94 | free( StartVbyte );
|
---|
95 | } // ~VbyteHeap
|
---|
96 |
|
---|
97 |
|
---|
98 | //######################### HandleNode #########################
|
---|
99 |
|
---|
100 |
|
---|
101 | // Create a handle node. The handle is not linked into the handle list. This is the responsibilitiy of the handle
|
---|
102 | // creator.
|
---|
103 |
|
---|
104 | void ?{}( HandleNode & this ) with(this) {
|
---|
105 | #ifdef VbyteDebug
|
---|
106 | serr | "enter:HandleNode::HandleNode, this:" | &this;
|
---|
107 | #endif // VbyteDebug
|
---|
108 | s = 0;
|
---|
109 | lnth = 0;
|
---|
110 | #ifdef VbyteDebug
|
---|
111 | serr | "exit:HandleNode::HandleNode, this:" | &this;
|
---|
112 | #endif // VbyteDebug
|
---|
113 | } // HandleNode
|
---|
114 |
|
---|
115 | // Create a handle node. The handle is linked into the handle list at the end. This means that this handle will NOT be
|
---|
116 | // in order by string address, but this is not a problem because a string with length zero does nothing during garbage
|
---|
117 | // collection.
|
---|
118 |
|
---|
119 | void ?{}( HandleNode & this, VbyteHeap & vh ) with(this) {
|
---|
120 | #ifdef VbyteDebug
|
---|
121 | serr | "enter:HandleNode::HandleNode, this:" | &this;
|
---|
122 | #endif // VbyteDebug
|
---|
123 | s = 0;
|
---|
124 | lnth = 0;
|
---|
125 | AddThisAfter( this, *vh.Header.blink );
|
---|
126 | #ifdef VbyteDebug
|
---|
127 | serr | "exit:HandleNode::HandleNode, this:" | &this;
|
---|
128 | #endif // VbyteDebug
|
---|
129 | } // HandleNode
|
---|
130 |
|
---|
131 |
|
---|
132 | // Delete a node from the handle list by unchaining it from the list. If the handle node was allocated dynamically, it
|
---|
133 | // is the responsibility of the creator to destroy it.
|
---|
134 |
|
---|
135 | void ^?{}( HandleNode & this ) with(this) {
|
---|
136 | #ifdef VbyteDebug
|
---|
137 | serr | "enter:HandleNode::~HandleNode, this:" | & this;
|
---|
138 | {
|
---|
139 | serr | nlOff;
|
---|
140 | serr | " lnth:" | lnth | " s:" | (void *)s | ",\"";
|
---|
141 | for ( int i = 0; i < lnth; i += 1 ) {
|
---|
142 | serr | s[i];
|
---|
143 | } // for
|
---|
144 | serr | "\" flink:" | flink | " blink:" | blink | nl;
|
---|
145 | serr | nlOn;
|
---|
146 | }
|
---|
147 | #endif // VbyteDebug
|
---|
148 | DeleteNode( this );
|
---|
149 | } // ~HandleNode
|
---|
150 |
|
---|
151 | //######################### String Resource #########################
|
---|
152 |
|
---|
153 |
|
---|
154 | VbyteHeap HeapArea;
|
---|
155 |
|
---|
156 | VbyteHeap * DEBUG_string_heap = & HeapArea;
|
---|
157 |
|
---|
158 | size_t DEBUG_string_bytes_avail_until_gc( VbyteHeap * heap ) {
|
---|
159 | return ((char*)heap->ExtVbyte) - heap->EndVbyte;
|
---|
160 | }
|
---|
161 |
|
---|
162 | const char * DEBUG_string_heap_start( VbyteHeap * heap ) {
|
---|
163 | return heap->StartVbyte;
|
---|
164 | }
|
---|
165 |
|
---|
166 |
|
---|
167 | // Returns the size of the string in bytes
|
---|
168 | size_t size(const string_res &s) with(s) {
|
---|
169 | return Handle.lnth;
|
---|
170 | }
|
---|
171 |
|
---|
172 | // Output operator
|
---|
173 | ofstream & ?|?(ofstream &out, const string_res &s) {
|
---|
174 | // Store auto-newline state so it can be restored
|
---|
175 | bool anl = getANL$(out);
|
---|
176 | nlOff(out);
|
---|
177 | for (size_t i = 0; i < s.Handle.lnth; i++) {
|
---|
178 | out | s[i];
|
---|
179 | }
|
---|
180 | out | sep;
|
---|
181 | // Re-apply newlines after done, for chaining version
|
---|
182 | if (anl) nlOn(out);
|
---|
183 | return out;
|
---|
184 | }
|
---|
185 |
|
---|
186 | void ?|?(ofstream &out, const string_res &s) {
|
---|
187 | // Store auto-newline state so it can be restored
|
---|
188 | bool anl = getANL$(out);
|
---|
189 | nlOff(out);
|
---|
190 | for (size_t i = 0; i < s.Handle.lnth; i++) {
|
---|
191 | // Need to re-apply on the last output operator, for whole-statement version
|
---|
192 | if (anl && i == s.Handle.lnth-1) nlOn(out);
|
---|
193 | out | s[i];
|
---|
194 | }
|
---|
195 | return out;
|
---|
196 | }
|
---|
197 |
|
---|
198 | // Empty constructor
|
---|
199 | void ?{}(string_res &s) with(s) {
|
---|
200 | (Handle){ HeapArea };
|
---|
201 | s.shareEditSet_prev = &s;
|
---|
202 | s.shareEditSet_next = &s;
|
---|
203 | }
|
---|
204 |
|
---|
205 | // Constructor from a raw buffer and size
|
---|
206 | void ?{}(string_res &s, const char* rhs, size_t rhslnth) with(s) {
|
---|
207 | (Handle){ HeapArea };
|
---|
208 | Handle.s = VbyteAlloc(HeapArea, rhslnth);
|
---|
209 | Handle.lnth = rhslnth;
|
---|
210 | for ( int i = 0; i < rhslnth; i += 1 ) { // copy characters
|
---|
211 | Handle.s[i] = rhs[i];
|
---|
212 | } // for
|
---|
213 | s.shareEditSet_prev = &s;
|
---|
214 | s.shareEditSet_next = &s;
|
---|
215 | }
|
---|
216 |
|
---|
217 | // String literal constructor
|
---|
218 | void ?{}(string_res &s, const char* rhs) {
|
---|
219 | (s){ rhs, strlen(rhs) };
|
---|
220 | }
|
---|
221 |
|
---|
222 | // General copy constructor
|
---|
223 | void ?{}(string_res &s, const string_res & s2, StrResInitMode mode, size_t start, size_t end ) {
|
---|
224 |
|
---|
225 | (s.Handle){ HeapArea };
|
---|
226 | s.Handle.s = s2.Handle.s + start;
|
---|
227 | s.Handle.lnth = end - start;
|
---|
228 | MoveThisAfter(s.Handle, s2.Handle ); // insert this handle after rhs handle
|
---|
229 | // ^ bug? skip others at early point in string
|
---|
230 |
|
---|
231 | if (mode == COPY_VALUE) {
|
---|
232 | // make s alone in its shareEditSet
|
---|
233 | s.shareEditSet_prev = &s;
|
---|
234 | s.shareEditSet_next = &s;
|
---|
235 | } else {
|
---|
236 | assert( mode == SHARE_EDITS );
|
---|
237 |
|
---|
238 | // s2 is logically const but not implementation const
|
---|
239 | string_res & s2mod = (string_res &) s2;
|
---|
240 |
|
---|
241 | // insert s after s2 on shareEditSet
|
---|
242 | s.shareEditSet_next = s2mod.shareEditSet_next;
|
---|
243 | s.shareEditSet_prev = &s2mod;
|
---|
244 | s.shareEditSet_next->shareEditSet_prev = &s;
|
---|
245 | s.shareEditSet_prev->shareEditSet_next = &s;
|
---|
246 | }
|
---|
247 | }
|
---|
248 |
|
---|
249 | void assign(string_res &this, const char* buffer, size_t bsize) {
|
---|
250 |
|
---|
251 | // traverse the incumbent share-edit set (SES) to recover the range of a base string to which `this` belongs
|
---|
252 | string_res * shareEditSetStartPeer = & this;
|
---|
253 | string_res * shareEditSetEndPeer = & this;
|
---|
254 | for (string_res * editPeer = this.shareEditSet_next; editPeer != &this; editPeer = editPeer->shareEditSet_next) {
|
---|
255 | if ( editPeer->Handle.s < shareEditSetStartPeer->Handle.s ) {
|
---|
256 | shareEditSetStartPeer = editPeer;
|
---|
257 | }
|
---|
258 | if ( shareEditSetEndPeer->Handle.s + shareEditSetEndPeer->Handle.lnth < editPeer->Handle.s + editPeer->Handle.lnth) {
|
---|
259 | shareEditSetEndPeer = editPeer;
|
---|
260 | }
|
---|
261 | }
|
---|
262 |
|
---|
263 | // full string is from start of shareEditSetStartPeer thru end of shareEditSetEndPeer
|
---|
264 | // `this` occurs in the middle of it, to be replaced
|
---|
265 | // build up the new text in `pasting`
|
---|
266 |
|
---|
267 | string_res pasting = {
|
---|
268 | shareEditSetStartPeer->Handle.s, // start of SES
|
---|
269 | this.Handle.s - shareEditSetStartPeer->Handle.s }; // length of SES, before this
|
---|
270 | append( pasting,
|
---|
271 | buffer, // start of replacement for this
|
---|
272 | bsize ); // length of replacement for this
|
---|
273 | append( pasting,
|
---|
274 | this.Handle.s + this.Handle.lnth, // start of SES after this
|
---|
275 | shareEditSetEndPeer->Handle.s + shareEditSetEndPeer->Handle.lnth -
|
---|
276 | (this.Handle.s + this.Handle.lnth) ); // length of SES, after this
|
---|
277 |
|
---|
278 | // The above string building can trigger compaction.
|
---|
279 | // The reference points (that are arguments of the string building) may move during that building.
|
---|
280 | // From this point on, they are stable.
|
---|
281 | // So now, capture their values for use in the overlap cases, below.
|
---|
282 | // Do not factor these definitions with the arguments used above.
|
---|
283 |
|
---|
284 | char * beforeBegin = shareEditSetStartPeer->Handle.s;
|
---|
285 | size_t beforeLen = this.Handle.s - beforeBegin;
|
---|
286 |
|
---|
287 | char * afterBegin = this.Handle.s + this.Handle.lnth;
|
---|
288 | size_t afterLen = shareEditSetEndPeer->Handle.s + shareEditSetEndPeer->Handle.lnth - afterBegin;
|
---|
289 |
|
---|
290 | size_t oldLnth = this.Handle.lnth;
|
---|
291 |
|
---|
292 | this.Handle.s = pasting.Handle.s + beforeLen;
|
---|
293 | this.Handle.lnth = bsize;
|
---|
294 | MoveThisAfter( this.Handle, pasting.Handle );
|
---|
295 |
|
---|
296 | // adjust all substring string and handle locations, and check if any substring strings are outside the new base string
|
---|
297 | char *limit = pasting.Handle.s + pasting.Handle.lnth;
|
---|
298 | for (string_res * p = this.shareEditSet_next; p != &this; p = p->shareEditSet_next) {
|
---|
299 | assert (p->Handle.s >= beforeBegin);
|
---|
300 | if ( p->Handle.s >= afterBegin ) {
|
---|
301 | assert ( p->Handle.s <= afterBegin + afterLen );
|
---|
302 | assert ( p->Handle.s + p->Handle.lnth <= afterBegin + afterLen );
|
---|
303 | // p starts after the edit
|
---|
304 | // take start and end as end-anchored
|
---|
305 | size_t startOffsetFromEnd = afterBegin + afterLen - p->Handle.s;
|
---|
306 | p->Handle.s = limit - startOffsetFromEnd;
|
---|
307 | // p->Handle.lnth unaffected
|
---|
308 | } else if ( p->Handle.s <= beforeBegin + beforeLen ) {
|
---|
309 | // p starts before, or at the start of, the edit
|
---|
310 | if ( p->Handle.s + p->Handle.lnth <= beforeBegin + beforeLen ) {
|
---|
311 | // p ends before the edit
|
---|
312 | // take end as start-anchored too
|
---|
313 | // p->Handle.lnth unaffected
|
---|
314 | } else if ( p->Handle.s + p->Handle.lnth < afterBegin ) {
|
---|
315 | // p ends during the edit; p does not include the last character replaced
|
---|
316 | // clip end of p to end at start of edit
|
---|
317 | p->Handle.lnth = beforeLen - ( p->Handle.s - beforeBegin );
|
---|
318 | } else {
|
---|
319 | // p ends after the edit
|
---|
320 | assert ( p->Handle.s + p->Handle.lnth <= afterBegin + afterLen );
|
---|
321 | // take end as end-anchored
|
---|
322 | // stretch-shrink p according to the edit
|
---|
323 | p->Handle.lnth += this.Handle.lnth;
|
---|
324 | p->Handle.lnth -= oldLnth;
|
---|
325 | }
|
---|
326 | // take start as start-anchored
|
---|
327 | size_t startOffsetFromStart = p->Handle.s - beforeBegin;
|
---|
328 | p->Handle.s = pasting.Handle.s + startOffsetFromStart;
|
---|
329 | } else {
|
---|
330 | assert ( p->Handle.s < afterBegin );
|
---|
331 | // p starts during the edit
|
---|
332 | assert( p->Handle.s + p->Handle.lnth >= beforeBegin + beforeLen );
|
---|
333 | if ( p->Handle.s + p->Handle.lnth < afterBegin ) {
|
---|
334 | // p ends during the edit; p does not include the last character replaced
|
---|
335 | // set p to empty string at start of edit
|
---|
336 | p->Handle.s = this.Handle.s;
|
---|
337 | p->Handle.lnth = 0;
|
---|
338 | } else {
|
---|
339 | // p includes the end of the edit
|
---|
340 | // clip start of p to start at end of edit
|
---|
341 | int charsToClip = afterBegin - p->Handle.s;
|
---|
342 | p->Handle.s = this.Handle.s + this.Handle.lnth;
|
---|
343 | p->Handle.lnth -= charsToClip;
|
---|
344 | }
|
---|
345 | }
|
---|
346 | MoveThisAfter( p->Handle, pasting.Handle ); // move substring handle to maintain sorted order by string position
|
---|
347 | }
|
---|
348 | }
|
---|
349 |
|
---|
350 | void ?=?(string_res &s, const char* other) {
|
---|
351 | assign(s, other, strlen(other));
|
---|
352 | }
|
---|
353 |
|
---|
354 | void ?=?(string_res &s, char other) {
|
---|
355 | assign(s, &other, 1);
|
---|
356 | }
|
---|
357 |
|
---|
358 | // Copy assignment operator
|
---|
359 | void ?=?(string_res & this, const string_res & rhs) with( this ) {
|
---|
360 | assign(this, rhs.Handle.s, rhs.Handle.lnth);
|
---|
361 | }
|
---|
362 |
|
---|
363 | void ?=?(string_res & this, string_res & rhs) with( this ) {
|
---|
364 | const string_res & rhs2 = rhs;
|
---|
365 | this = rhs2;
|
---|
366 | }
|
---|
367 |
|
---|
368 |
|
---|
369 | // Destructor
|
---|
370 | void ^?{}(string_res &s) with(s) {
|
---|
371 | // much delegated to implied ^VbyteSM
|
---|
372 |
|
---|
373 | // sever s from its share-edit peers, if any (four no-ops when already solo)
|
---|
374 | s.shareEditSet_prev->shareEditSet_next = s.shareEditSet_next;
|
---|
375 | s.shareEditSet_next->shareEditSet_prev = s.shareEditSet_prev;
|
---|
376 | s.shareEditSet_next = &s;
|
---|
377 | s.shareEditSet_prev = &s;
|
---|
378 | }
|
---|
379 |
|
---|
380 |
|
---|
381 | // Returns the character at the given index
|
---|
382 | // With unicode support, this may be different from just the byte at the given
|
---|
383 | // offset from the start of the string.
|
---|
384 | char ?[?](const string_res &s, size_t index) with(s) {
|
---|
385 | //TODO: Check if index is valid (no exceptions yet)
|
---|
386 | return Handle.s[index];
|
---|
387 | }
|
---|
388 |
|
---|
389 |
|
---|
390 | ///////////////////////////////////////////////////////////////////
|
---|
391 | // Concatenation
|
---|
392 |
|
---|
393 | void append(string_res &str1, const char * buffer, size_t bsize) {
|
---|
394 | size_t clnth = size(str1) + bsize;
|
---|
395 | if ( str1.Handle.s + size(str1) == buffer ) { // already juxtapose ?
|
---|
396 | // no-op
|
---|
397 | } else { // must copy some text
|
---|
398 | if ( str1.Handle.s + size(str1) == VbyteAlloc(HeapArea, 0) ) { // str1 at end of string area ?
|
---|
399 | VbyteAlloc(HeapArea, bsize); // create room for 2nd part at the end of string area
|
---|
400 | } else { // copy the two parts
|
---|
401 | char * str1oldBuf = str1.Handle.s;
|
---|
402 | str1.Handle.s = VbyteAlloc( HeapArea, clnth );
|
---|
403 | ByteCopy( HeapArea, str1.Handle.s, 0, str1.Handle.lnth, str1oldBuf, 0, str1.Handle.lnth);
|
---|
404 | } // if
|
---|
405 | ByteCopy( HeapArea, str1.Handle.s, str1.Handle.lnth, bsize, (char*)buffer, 0, (int)bsize);
|
---|
406 | // VbyteHeap & this, char *Dst, int DstStart, int DstLnth, char *Src, int SrcStart, int SrcLnth
|
---|
407 | } // if
|
---|
408 | str1.Handle.lnth = clnth;
|
---|
409 | }
|
---|
410 |
|
---|
411 | void ?+=?(string_res &str1, const string_res &str2) {
|
---|
412 | append( str1, str2.Handle.s, str2.Handle.lnth );
|
---|
413 | }
|
---|
414 |
|
---|
415 | void ?+=?(string_res &s, char other) {
|
---|
416 | append( s, &other, 1 );
|
---|
417 | }
|
---|
418 |
|
---|
419 | void ?+=?(string_res &s, const char* other) {
|
---|
420 | append( s, other, strlen(other) );
|
---|
421 | }
|
---|
422 |
|
---|
423 |
|
---|
424 |
|
---|
425 |
|
---|
426 | //////////////////////////////////////////////////////////
|
---|
427 | // Comparisons
|
---|
428 |
|
---|
429 |
|
---|
430 | bool ?==?(const string_res &s1, const string_res &s2) {
|
---|
431 | return ByteCmp( HeapArea, s1.Handle.s, 0, s1.Handle.lnth, s2.Handle.s, 0, s2.Handle.lnth) == 0;
|
---|
432 | }
|
---|
433 |
|
---|
434 | bool ?!=?(const string_res &s1, const string_res &s2) {
|
---|
435 | return !(s1 == s2);
|
---|
436 | }
|
---|
437 | bool ?==?(const string_res &s, const char* other) {
|
---|
438 | string_res sother = other;
|
---|
439 | return s == sother;
|
---|
440 | }
|
---|
441 | bool ?!=?(const string_res &s, const char* other) {
|
---|
442 | return !(s == other);
|
---|
443 | }
|
---|
444 |
|
---|
445 |
|
---|
446 | //////////////////////////////////////////////////////////
|
---|
447 | // Search
|
---|
448 |
|
---|
449 | bool contains(const string_res &s, char ch) {
|
---|
450 | for (i; size(s)) {
|
---|
451 | if (s[i] == ch) return true;
|
---|
452 | }
|
---|
453 | return false;
|
---|
454 | }
|
---|
455 |
|
---|
456 | int find(const string_res &s, char search) {
|
---|
457 | for (i; size(s)) {
|
---|
458 | if (s[i] == search) return i;
|
---|
459 | }
|
---|
460 | return size(s);
|
---|
461 | }
|
---|
462 |
|
---|
463 | /* Remaining implementations essentially ported from Sunjay's work */
|
---|
464 |
|
---|
465 | int find(const string_res &s, const string_res &search) {
|
---|
466 | return find(s, search.Handle.s, search.Handle.lnth);
|
---|
467 | }
|
---|
468 |
|
---|
469 | int find(const string_res &s, const char* search) {
|
---|
470 | return find(s, search, strlen(search));
|
---|
471 | }
|
---|
472 |
|
---|
473 | int find(const string_res &s, const char* search, size_t searchsize) {
|
---|
474 | // FIXME: This is a naive algorithm. We probably want to switch to someting
|
---|
475 | // like Boyer-Moore in the future.
|
---|
476 | // https://en.wikipedia.org/wiki/String_searching_algorithm
|
---|
477 |
|
---|
478 | // Always find the empty string
|
---|
479 | if (searchsize == 0) {
|
---|
480 | return 0;
|
---|
481 | }
|
---|
482 |
|
---|
483 | for (size_t i = 0; i < s.Handle.lnth; i++) {
|
---|
484 | size_t remaining = s.Handle.lnth - i;
|
---|
485 | // Never going to find the search string if the remaining string is
|
---|
486 | // smaller than search
|
---|
487 | if (remaining < searchsize) {
|
---|
488 | break;
|
---|
489 | }
|
---|
490 |
|
---|
491 | bool matched = true;
|
---|
492 | for (size_t j = 0; j < searchsize; j++) {
|
---|
493 | if (search[j] != s.Handle.s[i + j]) {
|
---|
494 | matched = false;
|
---|
495 | break;
|
---|
496 | }
|
---|
497 | }
|
---|
498 | if (matched) {
|
---|
499 | return i;
|
---|
500 | }
|
---|
501 | }
|
---|
502 |
|
---|
503 | return s.Handle.lnth;
|
---|
504 | }
|
---|
505 |
|
---|
506 | bool includes(const string_res &s, const string_res &search) {
|
---|
507 | return includes(s, search.Handle.s, search.Handle.lnth);
|
---|
508 | }
|
---|
509 |
|
---|
510 | bool includes(const string_res &s, const char* search) {
|
---|
511 | return includes(s, search, strlen(search));
|
---|
512 | }
|
---|
513 |
|
---|
514 | bool includes(const string_res &s, const char* search, size_t searchsize) {
|
---|
515 | return find(s, search, searchsize) < s.Handle.lnth;
|
---|
516 | }
|
---|
517 |
|
---|
518 | bool startsWith(const string_res &s, const string_res &prefix) {
|
---|
519 | return startsWith(s, prefix.Handle.s, prefix.Handle.lnth);
|
---|
520 | }
|
---|
521 |
|
---|
522 | bool startsWith(const string_res &s, const char* prefix) {
|
---|
523 | return startsWith(s, prefix, strlen(prefix));
|
---|
524 | }
|
---|
525 |
|
---|
526 | bool startsWith(const string_res &s, const char* prefix, size_t prefixsize) {
|
---|
527 | if (s.Handle.lnth < prefixsize) {
|
---|
528 | return false;
|
---|
529 | }
|
---|
530 | return memcmp(s.Handle.s, prefix, prefixsize) == 0;
|
---|
531 | }
|
---|
532 |
|
---|
533 | bool endsWith(const string_res &s, const string_res &suffix) {
|
---|
534 | return endsWith(s, suffix.Handle.s, suffix.Handle.lnth);
|
---|
535 | }
|
---|
536 |
|
---|
537 | bool endsWith(const string_res &s, const char* suffix) {
|
---|
538 | return endsWith(s, suffix, strlen(suffix));
|
---|
539 | }
|
---|
540 |
|
---|
541 | bool endsWith(const string_res &s, const char* suffix, size_t suffixsize) {
|
---|
542 | if (s.Handle.lnth < suffixsize) {
|
---|
543 | return false;
|
---|
544 | }
|
---|
545 | // Amount to offset the bytes pointer so that we are comparing the end of s
|
---|
546 | // to suffix. s.bytes + offset should be the first byte to compare against suffix
|
---|
547 | size_t offset = s.Handle.lnth - suffixsize;
|
---|
548 | return memcmp(s.Handle.s + offset, suffix, suffixsize) == 0;
|
---|
549 | }
|
---|
550 |
|
---|
551 | /* Back to Mike's work */
|
---|
552 |
|
---|
553 |
|
---|
554 | ///////////////////////////////////////////////////////////////////////////
|
---|
555 | // charclass, include, exclude
|
---|
556 |
|
---|
557 | void ?{}( charclass_res & this, const string_res & chars) {
|
---|
558 | (this){ chars.Handle.s, chars.Handle.lnth };
|
---|
559 | }
|
---|
560 |
|
---|
561 | void ?{}( charclass_res & this, const char * chars ) {
|
---|
562 | (this){ chars, strlen(chars) };
|
---|
563 | }
|
---|
564 |
|
---|
565 | void ?{}( charclass_res & this, const char * chars, size_t charssize ) {
|
---|
566 | (this.chars){ chars, charssize };
|
---|
567 | // now sort it ?
|
---|
568 | }
|
---|
569 |
|
---|
570 | void ^?{}( charclass_res & this ) {
|
---|
571 | ^(this.chars){};
|
---|
572 | }
|
---|
573 |
|
---|
574 | static bool test( const charclass_res & mask, char c ) {
|
---|
575 | // instead, use sorted char list?
|
---|
576 | return contains( mask.chars, c );
|
---|
577 | }
|
---|
578 |
|
---|
579 | int exclude(const string_res &s, const charclass_res &mask) {
|
---|
580 | for (int i = 0; i < size(s); i++) {
|
---|
581 | if ( test(mask, s[i]) ) return i;
|
---|
582 | }
|
---|
583 | return size(s);
|
---|
584 | }
|
---|
585 |
|
---|
586 | int include(const string_res &s, const charclass_res &mask) {
|
---|
587 | for (int i = 0; i < size(s); i++) {
|
---|
588 | if ( ! test(mask, s[i]) ) return i;
|
---|
589 | }
|
---|
590 | return size(s);
|
---|
591 | }
|
---|
592 |
|
---|
593 | //######################### VbyteHeap "implementation" #########################
|
---|
594 |
|
---|
595 |
|
---|
596 | // Add a new HandleNode node n after the current HandleNode node.
|
---|
597 |
|
---|
598 | static inline void AddThisAfter( HandleNode & this, HandleNode & n ) with(this) {
|
---|
599 | #ifdef VbyteDebug
|
---|
600 | serr | "enter:AddThisAfter, this:" | &this | " n:" | &n;
|
---|
601 | #endif // VbyteDebug
|
---|
602 | flink = n.flink;
|
---|
603 | blink = &n;
|
---|
604 | n.flink->blink = &this;
|
---|
605 | n.flink = &this;
|
---|
606 | #ifdef VbyteDebug
|
---|
607 | {
|
---|
608 | serr | "HandleList:";
|
---|
609 | serr | nlOff;
|
---|
610 | for ( HandleNode *ni = HeaderPtr->flink; ni != HeaderPtr; ni = ni->flink ) {
|
---|
611 | serr | "\tnode:" | ni | " lnth:" | ni->lnth | " s:" | (void *)ni->s | ",\"";
|
---|
612 | for ( int i = 0; i < ni->lnth; i += 1 ) {
|
---|
613 | serr | ni->s[i];
|
---|
614 | } // for
|
---|
615 | serr | "\" flink:" | ni->flink | " blink:" | ni->blink | nl;
|
---|
616 | } // for
|
---|
617 | serr | nlOn;
|
---|
618 | }
|
---|
619 | serr | "exit:AddThisAfter";
|
---|
620 | #endif // VbyteDebug
|
---|
621 | } // AddThisAfter
|
---|
622 |
|
---|
623 |
|
---|
624 | // Delete the current HandleNode node.
|
---|
625 |
|
---|
626 | static inline void DeleteNode( HandleNode & this ) with(this) {
|
---|
627 | #ifdef VbyteDebug
|
---|
628 | serr | "enter:DeleteNode, this:" | &this;
|
---|
629 | #endif // VbyteDebug
|
---|
630 | flink->blink = blink;
|
---|
631 | blink->flink = flink;
|
---|
632 | #ifdef VbyteDebug
|
---|
633 | serr | "exit:DeleteNode";
|
---|
634 | #endif // VbyteDebug
|
---|
635 | } // DeleteNode
|
---|
636 |
|
---|
637 |
|
---|
638 |
|
---|
639 | // Allocates specified storage for a string from byte-string area. If not enough space remains to perform the
|
---|
640 | // allocation, the garbage collection routine is called and a second attempt is made to allocate the space. If the
|
---|
641 | // second attempt fails, a further attempt is made to create a new, larger byte-string area.
|
---|
642 |
|
---|
643 | static inline char * VbyteAlloc( VbyteHeap & this, int size ) with(this) {
|
---|
644 | #ifdef VbyteDebug
|
---|
645 | serr | "enter:VbyteAlloc, size:" | size;
|
---|
646 | #endif // VbyteDebug
|
---|
647 | uintptr_t NoBytes;
|
---|
648 | char *r;
|
---|
649 |
|
---|
650 | NoBytes = ( uintptr_t )EndVbyte + size;
|
---|
651 | if ( NoBytes > ( uintptr_t )ExtVbyte ) { // enough room for new byte-string ?
|
---|
652 | garbage( this ); // firer up the garbage collector
|
---|
653 | NoBytes = ( uintptr_t )EndVbyte + size; // try again
|
---|
654 | if ( NoBytes > ( uintptr_t )ExtVbyte ) { // enough room for new byte-string ?
|
---|
655 | assert( 0 && "need to implement actual growth" );
|
---|
656 | // extend( size ); // extend the byte-string area
|
---|
657 | } // if
|
---|
658 | } // if
|
---|
659 | r = EndVbyte;
|
---|
660 | EndVbyte += size;
|
---|
661 | #ifdef VbyteDebug
|
---|
662 | serr | "exit:VbyteAlloc, r:" | (void *)r | " EndVbyte:" | (void *)EndVbyte | " ExtVbyte:" | ExtVbyte;
|
---|
663 | #endif // VbyteDebug
|
---|
664 | return r;
|
---|
665 | } // VbyteAlloc
|
---|
666 |
|
---|
667 |
|
---|
668 | // Move an existing HandleNode node h somewhere after the current HandleNode node so that it is in ascending order by
|
---|
669 | // the address in the byte string area.
|
---|
670 |
|
---|
671 | static inline void MoveThisAfter( HandleNode & this, const HandleNode & h ) with(this) {
|
---|
672 | #ifdef VbyteDebug
|
---|
673 | serr | "enter:MoveThisAfter, this:" | & this | " h:" | & h;
|
---|
674 | #endif // VbyteDebug
|
---|
675 | if ( s < h.s ) { // check argument values
|
---|
676 | // serr | "VbyteSM: Error - Cannot move byte string starting at:" | s | " after byte string starting at:"
|
---|
677 | // | ( h->s ) | " and keep handles in ascending order";
|
---|
678 | // exit(-1 );
|
---|
679 | assert( 0 && "VbyteSM: Error - Cannot move byte strings as requested and keep handles in ascending order");
|
---|
680 | } // if
|
---|
681 |
|
---|
682 | HandleNode *i;
|
---|
683 | for ( i = h.flink; i->s != 0 && s > ( i->s ); i = i->flink ); // find the position for this node after h
|
---|
684 | if ( & this != i->blink ) {
|
---|
685 | DeleteNode( this );
|
---|
686 | AddThisAfter( this, *i->blink );
|
---|
687 | } // if
|
---|
688 | #ifdef VbyteDebug
|
---|
689 | {
|
---|
690 | serr | "HandleList:";
|
---|
691 | serr | nlOff;
|
---|
692 | for ( HandleNode *n = HeaderPtr->flink; n != HeaderPtr; n = n->flink ) {
|
---|
693 | serr | "\tnode:" | n | " lnth:" | n->lnth | " s:" | (void *)n->s | ",\"";
|
---|
694 | for ( int i = 0; i < n->lnth; i += 1 ) {
|
---|
695 | serr | n->s[i];
|
---|
696 | } // for
|
---|
697 | serr | "\" flink:" | n->flink | " blink:" | n->blink | nl;
|
---|
698 | } // for
|
---|
699 | serr | nlOn;
|
---|
700 | }
|
---|
701 | serr | "exit:MoveThisAfter";
|
---|
702 | #endif // VbyteDebug
|
---|
703 | } // MoveThisAfter
|
---|
704 |
|
---|
705 |
|
---|
706 |
|
---|
707 |
|
---|
708 |
|
---|
709 | //######################### VbyteHeap #########################
|
---|
710 |
|
---|
711 | // Move characters from one location in the byte-string area to another. The routine handles the following situations:
|
---|
712 | //
|
---|
713 | // if the |Src| > |Dst| => truncate
|
---|
714 | // if the |Dst| > |Src| => pad Dst with blanks
|
---|
715 |
|
---|
716 | void ByteCopy( VbyteHeap & this, char *Dst, int DstStart, int DstLnth, char *Src, int SrcStart, int SrcLnth ) {
|
---|
717 | for ( int i = 0; i < DstLnth; i += 1 ) {
|
---|
718 | if ( i == SrcLnth ) { // |Dst| > |Src|
|
---|
719 | for ( ; i < DstLnth; i += 1 ) { // pad Dst with blanks
|
---|
720 | Dst[DstStart + i] = ' ';
|
---|
721 | } // for
|
---|
722 | break;
|
---|
723 | } // exit
|
---|
724 | Dst[DstStart + i] = Src[SrcStart + i];
|
---|
725 | } // for
|
---|
726 | } // ByteCopy
|
---|
727 |
|
---|
728 | // Compare two byte strings in the byte-string area. The routine returns the following values:
|
---|
729 | //
|
---|
730 | // 1 => Src1-byte-string > Src2-byte-string
|
---|
731 | // 0 => Src1-byte-string = Src2-byte-string
|
---|
732 | // -1 => Src1-byte-string < Src2-byte-string
|
---|
733 |
|
---|
734 | int ByteCmp( VbyteHeap & this, char *Src1, int Src1Start, int Src1Lnth, char *Src2, int Src2Start, int Src2Lnth ) with(this) {
|
---|
735 | #ifdef VbyteDebug
|
---|
736 | serr | "enter:ByteCmp, Src1Start:" | Src1Start | " Src1Lnth:" | Src1Lnth | " Src2Start:" | Src2Start | " Src2Lnth:" | Src2Lnth;
|
---|
737 | #endif // VbyteDebug
|
---|
738 | int cmp;
|
---|
739 |
|
---|
740 | CharZip: for ( int i = 0; ; i += 1 ) {
|
---|
741 | if ( i == Src2Lnth - 1 ) {
|
---|
742 | for ( ; ; i += 1 ) {
|
---|
743 | if ( i == Src1Lnth - 1 ) {
|
---|
744 | cmp = 0;
|
---|
745 | break CharZip;
|
---|
746 | } // exit
|
---|
747 | if ( Src1[Src1Start + i] != ' ') {
|
---|
748 | // SUSPECTED BUG: this could be be why Peter got the bug report about == " " (why is this case here at all?)
|
---|
749 | cmp = 1;
|
---|
750 | break CharZip;
|
---|
751 | } // exit
|
---|
752 | } // for
|
---|
753 | } // exit
|
---|
754 | if ( i == Src1Lnth - 1 ) {
|
---|
755 | for ( ; ; i += 1 ) {
|
---|
756 | if ( i == Src2Lnth - 1 ) {
|
---|
757 | cmp = 0;
|
---|
758 | break CharZip;
|
---|
759 | } // exit
|
---|
760 | if ( Src2[Src2Start + i] != ' ') {
|
---|
761 | cmp = -1;
|
---|
762 | break CharZip;
|
---|
763 | } // exit
|
---|
764 | } // for
|
---|
765 | } // exit
|
---|
766 | if ( Src2[Src2Start + i] != Src1[Src1Start+ i]) {
|
---|
767 | cmp = Src1[Src1Start + i] > Src2[Src2Start + i] ? 1 : -1;
|
---|
768 | break CharZip;
|
---|
769 | } // exit
|
---|
770 | } // for
|
---|
771 | #ifdef VbyteDebug
|
---|
772 | serr | "exit:ByteCmp, cmp:" | cmp;
|
---|
773 | #endif // VbyteDebug
|
---|
774 | return cmp;
|
---|
775 | } // ByteCmp
|
---|
776 |
|
---|
777 |
|
---|
778 | // The compaction moves all of the byte strings currently in use to the beginning of the byte-string area and modifies
|
---|
779 | // the handles to reflect the new positions of the byte strings. Compaction assumes that the handle list is in ascending
|
---|
780 | // order by pointers into the byte-string area. The strings associated with substrings do not have to be moved because
|
---|
781 | // the containing string has been moved. Hence, they only require that their string pointers be adjusted.
|
---|
782 |
|
---|
783 | void compaction(VbyteHeap & this) with(this) {
|
---|
784 | HandleNode *h;
|
---|
785 | char *obase, *nbase, *limit;
|
---|
786 |
|
---|
787 | NoOfCompactions += 1;
|
---|
788 | EndVbyte = StartVbyte;
|
---|
789 | h = Header.flink; // ignore header node
|
---|
790 | for (;;) {
|
---|
791 | ByteCopy( this, EndVbyte, 0, h->lnth, h->s, 0, h->lnth );
|
---|
792 | obase = h->s;
|
---|
793 | h->s = EndVbyte;
|
---|
794 | nbase = h->s;
|
---|
795 | EndVbyte += h->lnth;
|
---|
796 | limit = obase + h->lnth;
|
---|
797 | h = h->flink;
|
---|
798 |
|
---|
799 | // check if any substrings are allocated within a string
|
---|
800 |
|
---|
801 | for (;;) {
|
---|
802 | if ( h == &Header ) break; // end of header list ?
|
---|
803 | if ( h->s >= limit ) break; // outside of current string ?
|
---|
804 | h->s = nbase + (( uintptr_t )h->s - ( uintptr_t )obase );
|
---|
805 | h = h->flink;
|
---|
806 | } // for
|
---|
807 | if ( h == &Header ) break; // end of header list ?
|
---|
808 | } // for
|
---|
809 | } // compaction
|
---|
810 |
|
---|
811 |
|
---|
812 | // Garbage determines the amount of free space left in the heap and then reduces, leave the same, or extends the size of
|
---|
813 | // the heap. The heap is then compacted in the existing heap or into the newly allocated heap.
|
---|
814 |
|
---|
815 | void garbage(VbyteHeap & this ) with(this) {
|
---|
816 | #ifdef VbyteDebug
|
---|
817 | serr | "enter:garbage";
|
---|
818 | {
|
---|
819 | serr | "HandleList:";
|
---|
820 | for ( HandleNode *n = Header.flink; n != &Header; n = n->flink ) {
|
---|
821 | serr | nlOff;
|
---|
822 | serr | "\tnode:" | n | " lnth:" | n->lnth | " s:" | (void *)n->s | ",\"";
|
---|
823 | for ( int i = 0; i < n->lnth; i += 1 ) {
|
---|
824 | serr | n->s[i];
|
---|
825 | } // for
|
---|
826 | serr | nlOn;
|
---|
827 | serr | "\" flink:" | n->flink | " blink:" | n->blink;
|
---|
828 | } // for
|
---|
829 | }
|
---|
830 | #endif // VbyteDebug
|
---|
831 | int AmountUsed, AmountFree;
|
---|
832 |
|
---|
833 | AmountUsed = 0;
|
---|
834 | for ( HandleNode *i = Header.flink; i != &Header; i = i->flink ) { // calculate amount of byte area used
|
---|
835 | AmountUsed += i->lnth;
|
---|
836 | } // for
|
---|
837 | AmountFree = ( uintptr_t )ExtVbyte - ( uintptr_t )StartVbyte - AmountUsed;
|
---|
838 |
|
---|
839 | if ( AmountFree < ( int )( CurrSize * 0.1 )) { // free space less than 10% ?
|
---|
840 |
|
---|
841 | assert( 0 && "need to implement actual growth" );
|
---|
842 | // extend( CurrSize ); // extend the heap
|
---|
843 |
|
---|
844 | // Peter says, "This needs work before it should be used."
|
---|
845 | // } else if ( AmountFree > CurrSize / 2 ) { // free space greater than 3 times the initial allocation ?
|
---|
846 | // reduce(( AmountFree / CurrSize - 3 ) * CurrSize ); // reduce the memory
|
---|
847 |
|
---|
848 | } // if
|
---|
849 | compaction(this); // compact the byte area, in the same or new heap area
|
---|
850 | #ifdef VbyteDebug
|
---|
851 | {
|
---|
852 | serr | "HandleList:";
|
---|
853 | for ( HandleNode *n = Header.flink; n != &Header; n = n->flink ) {
|
---|
854 | serr | nlOff;
|
---|
855 | serr | "\tnode:" | n | " lnth:" | n->lnth | " s:" | (void *)n->s | ",\"";
|
---|
856 | for ( int i = 0; i < n->lnth; i += 1 ) {
|
---|
857 | serr | n->s[i];
|
---|
858 | } // for
|
---|
859 | serr | nlOn;
|
---|
860 | serr | "\" flink:" | n->flink | " blink:" | n->blink;
|
---|
861 | } // for
|
---|
862 | }
|
---|
863 | serr | "exit:garbage";
|
---|
864 | #endif // VbyteDebug
|
---|
865 | } // garbage
|
---|
866 |
|
---|
867 | #undef VbyteDebug
|
---|
868 |
|
---|
869 | //WIP
|
---|
870 | #if 0
|
---|
871 |
|
---|
872 |
|
---|
873 | // Extend the size of the byte-string area by creating a new area and copying the old area into it. The old byte-string
|
---|
874 | // area is deleted.
|
---|
875 |
|
---|
876 | void VbyteHeap::extend( int size ) {
|
---|
877 | #ifdef VbyteDebug
|
---|
878 | serr | "enter:extend, size:" | size;
|
---|
879 | #endif // VbyteDebug
|
---|
880 | char *OldStartVbyte;
|
---|
881 |
|
---|
882 | NoOfExtensions += 1;
|
---|
883 | OldStartVbyte = StartVbyte; // save previous byte area
|
---|
884 |
|
---|
885 | CurrSize += size > InitSize ? size : InitSize; // minimum extension, initial size
|
---|
886 | StartVbyte = EndVbyte = new char[CurrSize];
|
---|
887 | ExtVbyte = (void *)( StartVbyte + CurrSize );
|
---|
888 | compaction(); // copy from old heap to new & adjust pointers to new heap
|
---|
889 | delete OldStartVbyte; // release old heap
|
---|
890 | #ifdef VbyteDebug
|
---|
891 | serr | "exit:extend, CurrSize:" | CurrSize;
|
---|
892 | #endif // VbyteDebug
|
---|
893 | } // extend
|
---|
894 |
|
---|
895 |
|
---|
896 | // Extend the size of the byte-string area by creating a new area and copying the old area into it. The old byte-string
|
---|
897 | // area is deleted.
|
---|
898 |
|
---|
899 | void VbyteHeap::reduce( int size ) {
|
---|
900 | #ifdef VbyteDebug
|
---|
901 | serr | "enter:reduce, size:" | size;
|
---|
902 | #endif // VbyteDebug
|
---|
903 | char *OldStartVbyte;
|
---|
904 |
|
---|
905 | NoOfReductions += 1;
|
---|
906 | OldStartVbyte = StartVbyte; // save previous byte area
|
---|
907 |
|
---|
908 | CurrSize -= size;
|
---|
909 | StartVbyte = EndVbyte = new char[CurrSize];
|
---|
910 | ExtVbyte = (void *)( StartVbyte + CurrSize );
|
---|
911 | compaction(); // copy from old heap to new & adjust pointers to new heap
|
---|
912 | delete OldStartVbyte; // release old heap
|
---|
913 | #ifdef VbyteDebug
|
---|
914 | serr | "exit:reduce, CurrSize:" | CurrSize;
|
---|
915 | #endif // VbyteDebug
|
---|
916 | } // reduce
|
---|
917 |
|
---|
918 |
|
---|
919 | #endif
|
---|