1 // SciTE - Scintilla based Text Editor
\r
2 /** @file PropSet.cxx
\r
3 ** A Java style properties file module.
\r
5 // Copyright 1998-2003 by Neil Hodgson <neilh@scintilla.org>
\r
6 // The License.txt file describes the conditions under which this software may be distributed.
\r
8 // Maintain a dictionary of properties
\r
14 #include "Platform.h"
\r
16 #include "PropSet.h"
\r
18 #ifdef SCI_NAMESPACE
\r
19 using namespace Scintilla;
\r
22 // The comparison and case changing functions here assume ASCII
\r
23 // or extended ASCII such as the normal Windows code page.
\r
25 static inline char MakeUpperCase(char ch) {
\r
26 if (ch < 'a' || ch > 'z')
\r
29 return static_cast<char>(ch - 'a' + 'A');
\r
32 static inline bool IsLetter(char ch) {
\r
33 return ((ch >= 'a' && ch <= 'z') || (ch >= 'A' && ch <= 'Z'));
\r
36 inline bool IsASpace(unsigned int ch) {
\r
37 return (ch == ' ') || ((ch >= 0x09) && (ch <= 0x0d));
\r
40 int CompareCaseInsensitive(const char *a, const char *b) {
\r
43 char upperA = MakeUpperCase(*a);
\r
44 char upperB = MakeUpperCase(*b);
\r
45 if (upperA != upperB)
\r
46 return upperA - upperB;
\r
51 // Either *a or *b is nul
\r
55 int CompareNCaseInsensitive(const char *a, const char *b, size_t len) {
\r
56 while (*a && *b && len) {
\r
58 char upperA = MakeUpperCase(*a);
\r
59 char upperB = MakeUpperCase(*b);
\r
60 if (upperA != upperB)
\r
61 return upperA - upperB;
\r
70 // Either *a or *b is nul
\r
74 bool EqualCaseInsensitive(const char *a, const char *b) {
\r
75 return 0 == CompareCaseInsensitive(a, b);
\r
78 // Since the CaseInsensitive functions declared in SString
\r
79 // are implemented here, I will for now put the non-inline
\r
80 // implementations of the SString members here as well, so
\r
81 // that I can quickly see what effect this has.
\r
83 SString::SString(int i) : sizeGrowth(sizeGrowthDefault) {
\r
85 sprintf(number, "%0d", i);
\r
86 s = StringAllocate(number);
\r
87 sSize = sLen = (s) ? strlen(s) : 0;
\r
90 SString::SString(double d, int precision) : sizeGrowth(sizeGrowthDefault) {
\r
92 sprintf(number, "%.*f", precision, d);
\r
93 s = StringAllocate(number);
\r
94 sSize = sLen = (s) ? strlen(s) : 0;
\r
97 bool SString::grow(lenpos_t lenNew) {
\r
98 while (sizeGrowth * 6 < lenNew) {
\r
101 char *sNew = new char[lenNew + sizeGrowth + 1];
\r
104 memcpy(sNew, s, sLen);
\r
109 sSize = lenNew + sizeGrowth;
\r
114 SString &SString::assign(const char *sOther, lenpos_t sSize_) {
\r
117 } else if (sSize_ == measure_length) {
\r
118 sSize_ = strlen(sOther);
\r
120 if (sSize > 0 && sSize_ <= sSize) { // Does not allocate new buffer if the current is big enough
\r
122 memcpy(s, sOther, sSize_);
\r
128 s = StringAllocate(sOther, sSize_);
\r
130 sSize = sSize_; // Allow buffer bigger than real string, thus providing space to grow
\r
139 bool SString::operator==(const SString &sOther) const {
\r
140 if ((s == 0) && (sOther.s == 0))
\r
142 if ((s == 0) || (sOther.s == 0))
\r
144 return strcmp(s, sOther.s) == 0;
\r
147 bool SString::operator==(const char *sOther) const {
\r
148 if ((s == 0) && (sOther == 0))
\r
150 if ((s == 0) || (sOther == 0))
\r
152 return strcmp(s, sOther) == 0;
\r
155 SString SString::substr(lenpos_t subPos, lenpos_t subLen) const {
\r
156 if (subPos >= sLen) {
\r
157 return SString(); // return a null string if start index is out of bounds
\r
159 if ((subLen == measure_length) || (subPos + subLen > sLen)) {
\r
160 subLen = sLen - subPos; // can't substr past end of source string
\r
162 return SString(s, subPos, subPos + subLen);
\r
165 SString &SString::lowercase(lenpos_t subPos, lenpos_t subLen) {
\r
166 if ((subLen == measure_length) || (subPos + subLen > sLen)) {
\r
167 subLen = sLen - subPos; // don't apply past end of string
\r
169 for (lenpos_t i = subPos; i < subPos + subLen; i++) {
\r
170 if (s[i] < 'A' || s[i] > 'Z')
\r
173 s[i] = static_cast<char>(s[i] - 'A' + 'a');
\r
178 SString &SString::uppercase(lenpos_t subPos, lenpos_t subLen) {
\r
179 if ((subLen == measure_length) || (subPos + subLen > sLen)) {
\r
180 subLen = sLen - subPos; // don't apply past end of string
\r
182 for (lenpos_t i = subPos; i < subPos + subLen; i++) {
\r
183 if (s[i] < 'a' || s[i] > 'z')
\r
186 s[i] = static_cast<char>(s[i] - 'a' + 'A');
\r
191 SString &SString::append(const char *sOther, lenpos_t sLenOther, char sep) {
\r
195 if (sLenOther == measure_length) {
\r
196 sLenOther = strlen(sOther);
\r
199 if (sLen && sep) { // Only add a separator if not empty
\r
202 lenpos_t lenNew = sLen + sLenOther + lenSep;
\r
203 // Conservative about growing the buffer: don't do it, unless really needed
\r
204 if ((lenNew < sSize) || (grow(lenNew))) {
\r
209 memcpy(&s[sLen], sOther, sLenOther);
\r
216 SString &SString::insert(lenpos_t pos, const char *sOther, lenpos_t sLenOther) {
\r
217 if (!sOther || pos > sLen) {
\r
220 if (sLenOther == measure_length) {
\r
221 sLenOther = strlen(sOther);
\r
223 lenpos_t lenNew = sLen + sLenOther;
\r
224 // Conservative about growing the buffer: don't do it, unless really needed
\r
225 if ((lenNew < sSize) || grow(lenNew)) {
\r
226 lenpos_t moveChars = sLen - pos + 1;
\r
227 for (lenpos_t i = moveChars; i > 0; i--) {
\r
228 s[pos + sLenOther + i - 1] = s[pos + i - 1];
\r
230 memcpy(s + pos, sOther, sLenOther);
\r
237 * Remove @a len characters from the @a pos position, included.
\r
238 * Characters at pos + len and beyond replace characters at pos.
\r
239 * If @a len is 0, or greater than the length of the string
\r
240 * starting at @a pos, the string is just truncated at @a pos.
\r
242 void SString::remove(lenpos_t pos, lenpos_t len) {
\r
246 if (len < 1 || pos + len >= sLen) {
\r
250 for (lenpos_t i = pos; i < sLen - len + 1; i++) {
\r
257 bool SString::startswith(const char *prefix) {
\r
258 lenpos_t lenPrefix = strlen(prefix);
\r
259 if (lenPrefix > sLen) {
\r
262 return strncmp(s, prefix, lenPrefix) == 0;
\r
265 bool SString::endswith(const char *suffix) {
\r
266 lenpos_t lenSuffix = strlen(suffix);
\r
267 if (lenSuffix > sLen) {
\r
270 return strncmp(s + sLen - lenSuffix, suffix, lenSuffix) == 0;
\r
273 int SString::search(const char *sFind, lenpos_t start) const {
\r
274 if (start < sLen) {
\r
275 const char *sFound = strstr(s + start, sFind);
\r
283 int SString::substitute(char chFind, char chReplace) {
\r
287 t = strchr(t, chFind);
\r
297 int SString::substitute(const char *sFind, const char *sReplace) {
\r
299 lenpos_t lenFind = strlen(sFind);
\r
300 lenpos_t lenReplace = strlen(sReplace);
\r
301 int posFound = search(sFind);
\r
302 while (posFound >= 0) {
\r
303 remove(posFound, lenFind);
\r
304 insert(posFound, sReplace, lenReplace);
\r
305 posFound = search(sFind, posFound + lenReplace);
\r
311 char *SContainer::StringAllocate(lenpos_t len) {
\r
312 if (len != measure_length) {
\r
313 return new char[len + 1];
\r
319 char *SContainer::StringAllocate(const char *s, lenpos_t len) {
\r
323 if (len == measure_length) {
\r
326 char *sNew = new char[len + 1];
\r
328 memcpy(sNew, s, len);
\r
334 // End SString functions
\r
336 PropSet::PropSet() {
\r
338 for (int root = 0; root < hashRoots; root++)
\r
342 PropSet::~PropSet() {
\r
347 void PropSet::Set(const char *key, const char *val, int lenKey, int lenVal) {
\r
348 if (!*key) // Empty keys are not supported
\r
351 lenKey = static_cast<int>(strlen(key));
\r
353 lenVal = static_cast<int>(strlen(val));
\r
354 unsigned int hash = HashString(key, lenKey);
\r
355 for (Property *p = props[hash % hashRoots]; p; p = p->next) {
\r
356 if ((hash == p->hash) &&
\r
357 ((strlen(p->key) == static_cast<unsigned int>(lenKey)) &&
\r
358 (0 == strncmp(p->key, key, lenKey)))) {
\r
359 // Replace current value
\r
361 p->val = StringDup(val, lenVal);
\r
366 Property *pNew = new Property;
\r
369 pNew->key = StringDup(key, lenKey);
\r
370 pNew->val = StringDup(val, lenVal);
\r
371 pNew->next = props[hash % hashRoots];
\r
372 props[hash % hashRoots] = pNew;
\r
376 void PropSet::Set(const char *keyVal) {
\r
377 while (IsASpace(*keyVal))
\r
379 const char *endVal = keyVal;
\r
380 while (*endVal && (*endVal != '\n'))
\r
382 const char *eqAt = strchr(keyVal, '=');
\r
384 Set(keyVal, eqAt + 1, eqAt-keyVal, endVal - eqAt - 1);
\r
385 } else if (*keyVal) { // No '=' so assume '=1'
\r
386 Set(keyVal, "1", endVal-keyVal, 1);
\r
390 void PropSet::Unset(const char *key, int lenKey) {
\r
391 if (!*key) // Empty keys are not supported
\r
394 lenKey = static_cast<int>(strlen(key));
\r
395 unsigned int hash = HashString(key, lenKey);
\r
396 Property *pPrev = NULL;
\r
397 for (Property *p = props[hash % hashRoots]; p; p = p->next) {
\r
398 if ((hash == p->hash) &&
\r
399 ((strlen(p->key) == static_cast<unsigned int>(lenKey)) &&
\r
400 (0 == strncmp(p->key, key, lenKey)))) {
\r
402 pPrev->next = p->next;
\r
404 props[hash % hashRoots] = p->next;
\r
406 enumnext = p->next; // Not that anyone should mix enum and Set / Unset.
\r
417 void PropSet::SetMultiple(const char *s) {
\r
418 const char *eol = strchr(s, '\n');
\r
422 eol = strchr(s, '\n');
\r
427 SString PropSet::Get(const char *key) const {
\r
428 unsigned int hash = HashString(key, strlen(key));
\r
429 for (Property *p = props[hash % hashRoots]; p; p = p->next) {
\r
430 if ((hash == p->hash) && (0 == strcmp(p->key, key))) {
\r
435 // Failed here, so try in base property set
\r
436 return superPS->Get(key);
\r
442 // There is some inconsistency between GetExpanded("foo") and Expand("$(foo)").
\r
443 // A solution is to keep a stack of variables that have been expanded, so that
\r
444 // recursive expansions can be skipped. For now I'll just use the C++ stack
\r
445 // for that, through a recursive function and a simple chain of pointers.
\r
448 VarChain(const char*var_=NULL, const VarChain *link_=NULL): var(var_), link(link_) {}
\r
450 bool contains(const char *testVar) const {
\r
451 return (var && (0 == strcmp(var, testVar)))
\r
452 || (link && link->contains(testVar));
\r
456 const VarChain *link;
\r
459 static int ExpandAllInPlace(const PropSet &props, SString &withVars, int maxExpands, const VarChain &blankVars = VarChain()) {
\r
460 int varStart = withVars.search("$(");
\r
461 while ((varStart >= 0) && (maxExpands > 0)) {
\r
462 int varEnd = withVars.search(")", varStart+2);
\r
467 // For consistency, when we see '$(ab$(cde))', expand the inner variable first,
\r
468 // regardless whether there is actually a degenerate variable named 'ab$(cde'.
\r
469 int innerVarStart = withVars.search("$(", varStart+2);
\r
470 while ((innerVarStart > varStart) && (innerVarStart < varEnd)) {
\r
471 varStart = innerVarStart;
\r
472 innerVarStart = withVars.search("$(", varStart+2);
\r
475 SString var(withVars.c_str(), varStart + 2, varEnd);
\r
476 SString val = props.Get(var.c_str());
\r
478 if (blankVars.contains(var.c_str())) {
\r
479 val.clear(); // treat blankVar as an empty string (e.g. to block self-reference)
\r
482 if (--maxExpands >= 0) {
\r
483 maxExpands = ExpandAllInPlace(props, val, maxExpands, VarChain(var.c_str(), &blankVars));
\r
486 withVars.remove(varStart, varEnd-varStart+1);
\r
487 withVars.insert(varStart, val.c_str(), val.length());
\r
489 varStart = withVars.search("$(");
\r
495 SString PropSet::GetExpanded(const char *key) const {
\r
496 SString val = Get(key);
\r
497 ExpandAllInPlace(*this, val, 100, VarChain(key));
\r
501 SString PropSet::Expand(const char *withVars, int maxExpands) const {
\r
502 SString val = withVars;
\r
503 ExpandAllInPlace(*this, val, maxExpands);
\r
507 int PropSet::GetInt(const char *key, int defaultValue) const {
\r
508 SString val = GetExpanded(key);
\r
510 return val.value();
\r
511 return defaultValue;
\r
514 bool isprefix(const char *target, const char *prefix) {
\r
515 while (*target && *prefix) {
\r
516 if (*target != *prefix)
\r
527 void PropSet::Clear() {
\r
528 for (int root = 0; root < hashRoots; root++) {
\r
529 Property *p = props[root];
\r
531 Property *pNext = p->next;
\r
544 char *PropSet::ToString() const {
\r
546 for (int r = 0; r < hashRoots; r++) {
\r
547 for (Property *p = props[r]; p; p = p->next) {
\r
548 len += strlen(p->key) + 1;
\r
549 len += strlen(p->val) + 1;
\r
553 len = 1; // Return as empty string
\r
554 char *ret = new char [len];
\r
557 for (int root = 0; root < hashRoots; root++) {
\r
558 for (Property *p = props[root]; p; p = p->next) {
\r
560 w += strlen(p->key);
\r
563 w += strlen(p->val);
\r
573 * Creates an array that points into each word in the string and puts \0 terminators
\r
576 static char **ArrayFromWordList(char *wordlist, int *len, bool onlyLineEnds = false) {
\r
579 // For rapid determination of whether a character is a separator, build
\r
580 // a look up table.
\r
581 bool wordSeparator[256];
\r
582 for (int i=0;i<256; i++) {
\r
583 wordSeparator[i] = false;
\r
585 wordSeparator['\r'] = true;
\r
586 wordSeparator['\n'] = true;
\r
587 if (!onlyLineEnds) {
\r
588 wordSeparator[' '] = true;
\r
589 wordSeparator['\t'] = true;
\r
591 for (int j = 0; wordlist[j]; j++) {
\r
592 int curr = static_cast<unsigned char>(wordlist[j]);
\r
593 if (!wordSeparator[curr] && wordSeparator[prev])
\r
597 char **keywords = new char *[words + 1];
\r
601 size_t slen = strlen(wordlist);
\r
602 for (size_t k = 0; k < slen; k++) {
\r
603 if (!wordSeparator[static_cast<unsigned char>(wordlist[k])]) {
\r
605 keywords[words] = &wordlist[k];
\r
609 wordlist[k] = '\0';
\r
611 prev = wordlist[k];
\r
613 keywords[words] = &wordlist[slen];
\r
621 void WordList::Clear() {
\r
632 void WordList::Set(const char *s) {
\r
633 list = StringDup(s);
\r
635 words = ArrayFromWordList(list, &len, onlyLineEnds);
\r
638 extern "C" int cmpString(const void *a1, const void *a2) {
\r
639 // Can't work out the correct incantation to use modern casts here
\r
640 return strcmp(*(char**)(a1), *(char**)(a2));
\r
643 static void SortWordList(char **words, unsigned int len) {
\r
644 qsort(reinterpret_cast<void*>(words), len, sizeof(*words),
\r
648 bool WordList::InList(const char *s) {
\r
653 SortWordList(words, len);
\r
654 for (unsigned int k = 0; k < (sizeof(starts) / sizeof(starts[0])); k++)
\r
656 for (int l = len - 1; l >= 0; l--) {
\r
657 unsigned char indexChar = words[l][0];
\r
658 starts[indexChar] = l;
\r
661 unsigned char firstChar = s[0];
\r
662 int j = starts[firstChar];
\r
664 while ((unsigned char)words[j][0] == firstChar) {
\r
665 if (s[1] == words[j][1]) {
\r
666 const char *a = words[j] + 1;
\r
667 const char *b = s + 1;
\r
668 while (*a && *a == *b) {
\r
680 while (words[j][0] == '^') {
\r
681 const char *a = words[j] + 1;
\r
683 while (*a && *a == *b) {
\r
695 /** similar to InList, but word s can be a substring of keyword.
\r
696 * eg. the keyword define is defined as def~ine. This means the word must start
\r
697 * with def to be a keyword, but also defi, defin and define are valid.
\r
698 * The marker is ~ in this case.
\r
700 bool WordList::InListAbbreviated(const char *s, const char marker) {
\r
705 SortWordList(words, len);
\r
706 for (unsigned int k = 0; k < (sizeof(starts) / sizeof(starts[0])); k++)
\r
708 for (int l = len - 1; l >= 0; l--) {
\r
709 unsigned char indexChar = words[l][0];
\r
710 starts[indexChar] = l;
\r
713 unsigned char firstChar = s[0];
\r
714 int j = starts[firstChar];
\r
716 while (words[j][0] == firstChar) {
\r
717 bool isSubword = false;
\r
719 if (words[j][1] == marker) {
\r
723 if (s[1] == words[j][start]) {
\r
724 const char *a = words[j] + start;
\r
725 const char *b = s + 1;
\r
726 while (*a && *a == *b) {
\r
728 if (*a == marker) {
\r
734 if ((!*a || isSubword) && !*b)
\r
742 while (words[j][0] == '^') {
\r
743 const char *a = words[j] + 1;
\r
745 while (*a && *a == *b) {
\r