|
a |
|
b/cfgui/picoxml.h |
|
|
1 |
/* Copyright (C) 2016 J.F.Dockes
|
|
|
2 |
* This program is free software; you can redistribute it and/or modify
|
|
|
3 |
* it under the terms of the GNU General Public License as published by
|
|
|
4 |
* the Free Software Foundation; either version 2 of the License, or
|
|
|
5 |
* (at your option) any later version.
|
|
|
6 |
*
|
|
|
7 |
* This program is distributed in the hope that it will be useful,
|
|
|
8 |
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
9 |
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
10 |
* GNU General Public License for more details.
|
|
|
11 |
*
|
|
|
12 |
* You should have received a copy of the GNU General Public License
|
|
|
13 |
* along with this program; if not, write to the
|
|
|
14 |
* Free Software Foundation, Inc.,
|
|
|
15 |
* 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
|
|
|
16 |
*/
|
|
|
17 |
|
|
|
18 |
#ifndef _PICOXML_H_INCLUDED_
|
|
|
19 |
#define _PICOXML_H_INCLUDED_
|
|
|
20 |
|
|
|
21 |
/**
|
|
|
22 |
* PicoXMLParser: a single include file parser for an XML-like, but
|
|
|
23 |
* restricted language, adequate for config files, not for arbitrary
|
|
|
24 |
* externally generated data.
|
|
|
25 |
*
|
|
|
26 |
* - The code depends on nothing but the "classical" C++ standard
|
|
|
27 |
* library (c++11 not necessary).
|
|
|
28 |
* - The input to the parser is a single c++ string. Does not deal with
|
|
|
29 |
* input in several pieces or files.
|
|
|
30 |
* - SAX mode only. You have access to the tag stack. I've always
|
|
|
31 |
* found DOM mode less usable.
|
|
|
32 |
* - Checks for proper tag nesting and not much else.
|
|
|
33 |
* - ! No CDATA
|
|
|
34 |
* - ! Attributes should really really not contain XML special chars.
|
|
|
35 |
* - Entity decoding is left as an exercise to the user.
|
|
|
36 |
*
|
|
|
37 |
* A typical input would be like the following (you can add XML
|
|
|
38 |
* declarations, whitespace and newlines to taste).
|
|
|
39 |
*
|
|
|
40 |
* <top>top chrs1<sub attr="attrval">sub chrs</sub>top chrs2 <emptyelt /></top>
|
|
|
41 |
*
|
|
|
42 |
* Usage: subclass PicoXMLParser, overriding the methods in the
|
|
|
43 |
* "protected:" section (look there for more details), call the
|
|
|
44 |
* constructor with your input, then call parse().
|
|
|
45 |
*/
|
|
|
46 |
|
|
|
47 |
#include <string>
|
|
|
48 |
#include <vector>
|
|
|
49 |
#include <map>
|
|
|
50 |
#include <sstream>
|
|
|
51 |
#include <iostream>
|
|
|
52 |
|
|
|
53 |
class PicoXMLParser {
|
|
|
54 |
public:
|
|
|
55 |
PicoXMLParser(const std::string& input)
|
|
|
56 |
: m_in(input), m_pos(0) {
|
|
|
57 |
}
|
|
|
58 |
virtual ~PicoXMLParser() { }
|
|
|
59 |
|
|
|
60 |
virtual bool parse() {
|
|
|
61 |
// skip initial whitespace and XML decl. On success, returns with
|
|
|
62 |
// current pos on first tag '<'
|
|
|
63 |
if (!skipDecl()) {
|
|
|
64 |
return false;
|
|
|
65 |
}
|
|
|
66 |
if (nomore()) {
|
|
|
67 |
// empty file
|
|
|
68 |
return true;
|
|
|
69 |
}
|
|
|
70 |
|
|
|
71 |
for (;;) {
|
|
|
72 |
// Current char is '<' and the next char is not '?'
|
|
|
73 |
//std::cerr<< "m_pos "<< m_pos<<" char "<< m_in[m_pos] << std::endl;
|
|
|
74 |
m_pos++;
|
|
|
75 |
if (nomore()) {
|
|
|
76 |
m_reason << "EOF within tag";
|
|
|
77 |
return false;
|
|
|
78 |
}
|
|
|
79 |
std::string::size_type spos = m_pos;
|
|
|
80 |
int isendtag = m_in[m_pos] == '/' ? 1 : 0;
|
|
|
81 |
|
|
|
82 |
skipStr(">");
|
|
|
83 |
if (m_pos == std::string::npos || m_pos <= spos + 1) {
|
|
|
84 |
m_reason << "Empty tag or EOF inside tag. pos " << spos;
|
|
|
85 |
return false;
|
|
|
86 |
}
|
|
|
87 |
|
|
|
88 |
int emptyel = m_in[m_pos-2] == '/' ? 1 : 0;
|
|
|
89 |
if (emptyel && isendtag) {
|
|
|
90 |
m_reason << "Bad tag </xx/> at cpos " << spos;
|
|
|
91 |
return false;
|
|
|
92 |
}
|
|
|
93 |
|
|
|
94 |
std::string tag =
|
|
|
95 |
m_in.substr(spos + isendtag,
|
|
|
96 |
m_pos - (spos + 1 + isendtag + emptyel));
|
|
|
97 |
//std::cerr << "TAG NAME [" << tag << "]\n";
|
|
|
98 |
trimtag(tag);
|
|
|
99 |
std::map<std::string, std::string> attrs;
|
|
|
100 |
if (!parseattrs(tag, attrs)) {
|
|
|
101 |
return false;
|
|
|
102 |
}
|
|
|
103 |
if (isendtag) {
|
|
|
104 |
if (m_tagstack.empty() || tag.compare(m_tagstack.back())) {
|
|
|
105 |
m_reason << "Closing not open tag " << tag <<
|
|
|
106 |
" at cpos " << m_pos;
|
|
|
107 |
return false;
|
|
|
108 |
}
|
|
|
109 |
m_tagstack.pop_back();
|
|
|
110 |
endElement(tag);
|
|
|
111 |
} else {
|
|
|
112 |
startElement(tag, attrs);
|
|
|
113 |
m_tagstack.push_back(tag);
|
|
|
114 |
if (emptyel) {
|
|
|
115 |
m_tagstack.pop_back();
|
|
|
116 |
endElement(tag);
|
|
|
117 |
}
|
|
|
118 |
}
|
|
|
119 |
spos = m_pos;
|
|
|
120 |
m_pos = m_in.find("<", m_pos);
|
|
|
121 |
if (nomore()) {
|
|
|
122 |
if (!m_tagstack.empty()) {
|
|
|
123 |
m_reason << "EOF hit inside open element";
|
|
|
124 |
return false;
|
|
|
125 |
}
|
|
|
126 |
return true;
|
|
|
127 |
}
|
|
|
128 |
if (m_pos != spos) {
|
|
|
129 |
characterData(m_in.substr(spos, m_pos - spos));
|
|
|
130 |
}
|
|
|
131 |
}
|
|
|
132 |
return false;
|
|
|
133 |
}
|
|
|
134 |
|
|
|
135 |
virtual std::string getReason() {
|
|
|
136 |
return m_reason.str();
|
|
|
137 |
}
|
|
|
138 |
|
|
|
139 |
protected:
|
|
|
140 |
|
|
|
141 |
/* Methods to be overriden */
|
|
|
142 |
|
|
|
143 |
/**
|
|
|
144 |
* Called when seeing an opening tag.
|
|
|
145 |
* @param tagname the tag name
|
|
|
146 |
* @param attrs a map of attribute name/value pairs
|
|
|
147 |
*/
|
|
|
148 |
virtual void startElement(const std::string& /*tagname*/,
|
|
|
149 |
const std::map<std::string, std::string>&
|
|
|
150 |
/* attrs */) {
|
|
|
151 |
}
|
|
|
152 |
|
|
|
153 |
/**
|
|
|
154 |
* Called when closing a tag. You should probably have been
|
|
|
155 |
* accumulating text and stuff since the tag opening.
|
|
|
156 |
* @param tagname the tag name.
|
|
|
157 |
*/
|
|
|
158 |
virtual void endElement(const std::string& /*tagname*/) {}
|
|
|
159 |
|
|
|
160 |
/*
|
|
|
161 |
* Called when we see non-tag data.
|
|
|
162 |
* @param data the data.
|
|
|
163 |
*/
|
|
|
164 |
virtual void characterData(const std::string& /*data*/) {}
|
|
|
165 |
|
|
|
166 |
/*
|
|
|
167 |
* Gives access to the current path in the tree. Attributes are
|
|
|
168 |
* not kept in there though, you'll have to do this yourself.
|
|
|
169 |
* @return a const ref to a vector of tag names.
|
|
|
170 |
*/
|
|
|
171 |
virtual const std::vector<std::string>& tagStack() {
|
|
|
172 |
return m_tagstack;
|
|
|
173 |
}
|
|
|
174 |
|
|
|
175 |
|
|
|
176 |
private:
|
|
|
177 |
|
|
|
178 |
const std::string& m_in;
|
|
|
179 |
std::string::size_type m_pos;
|
|
|
180 |
std::stringstream m_reason;
|
|
|
181 |
std::vector<std::string> m_tagstack;
|
|
|
182 |
|
|
|
183 |
bool nomore() const {
|
|
|
184 |
return m_pos == std::string::npos || m_pos == m_in.size();
|
|
|
185 |
}
|
|
|
186 |
bool skipWS(const std::string& in, std::string::size_type& pos) {
|
|
|
187 |
if (pos == std::string::npos)
|
|
|
188 |
return false;
|
|
|
189 |
pos = in.find_first_not_of(" \t\n\r", pos);
|
|
|
190 |
return pos != std::string::npos;
|
|
|
191 |
}
|
|
|
192 |
bool skipStr(const std::string& str) {
|
|
|
193 |
if (m_pos == std::string::npos)
|
|
|
194 |
return false;
|
|
|
195 |
m_pos = m_in.find(str, m_pos);
|
|
|
196 |
if (m_pos != std::string::npos)
|
|
|
197 |
m_pos += str.size();
|
|
|
198 |
return m_pos != std::string::npos;
|
|
|
199 |
}
|
|
|
200 |
int peek() const {
|
|
|
201 |
if (nomore())
|
|
|
202 |
return -1;
|
|
|
203 |
return m_in[m_pos + 1];
|
|
|
204 |
}
|
|
|
205 |
void trimtag(std::string& tagname) {
|
|
|
206 |
std::string::size_type trimpos = tagname.find_last_not_of(" \t\n\r");
|
|
|
207 |
if (trimpos != std::string::npos) {
|
|
|
208 |
tagname = tagname.substr(0, trimpos+1);
|
|
|
209 |
}
|
|
|
210 |
}
|
|
|
211 |
|
|
|
212 |
bool skipDecl() {
|
|
|
213 |
for (;;) {
|
|
|
214 |
if (!skipWS(m_in, m_pos)) {
|
|
|
215 |
m_reason << "EOF during initial ws skip";
|
|
|
216 |
return true;
|
|
|
217 |
}
|
|
|
218 |
if (m_in[m_pos] != '<') {
|
|
|
219 |
m_reason << "EOF file does not begin with decl/tag";
|
|
|
220 |
return false;
|
|
|
221 |
}
|
|
|
222 |
if (peek() == '?') {
|
|
|
223 |
if (!skipStr("?>")) {
|
|
|
224 |
m_reason << "EOF while looking for end of xml decl";
|
|
|
225 |
return false;
|
|
|
226 |
}
|
|
|
227 |
} else {
|
|
|
228 |
break;
|
|
|
229 |
}
|
|
|
230 |
}
|
|
|
231 |
return true;
|
|
|
232 |
}
|
|
|
233 |
|
|
|
234 |
bool parseattrs(std::string& tag,
|
|
|
235 |
std::map<std::string, std::string>& attrs) {
|
|
|
236 |
//std::cerr << "parseattrs: [" << tag << "]\n";
|
|
|
237 |
attrs.clear();
|
|
|
238 |
std::string::size_type spos = tag.find_first_of(" \t\n\r");
|
|
|
239 |
if (spos == std::string::npos)
|
|
|
240 |
return true;
|
|
|
241 |
std::string tagname = tag.substr(0, spos);
|
|
|
242 |
//std::cerr << "tag name [" << tagname << "] pos " << spos << "\n";
|
|
|
243 |
skipWS(tag, spos);
|
|
|
244 |
|
|
|
245 |
for (;;) {
|
|
|
246 |
//std::cerr << "top of loop [" << tag.substr(spos) << "]\n";
|
|
|
247 |
std::string::size_type epos = tag.find_first_of(" \t\n\r=", spos);
|
|
|
248 |
if (epos == std::string::npos) {
|
|
|
249 |
m_reason << "Bad attributes syntax at cpos " << m_pos + epos;
|
|
|
250 |
return false;
|
|
|
251 |
}
|
|
|
252 |
std::string attrnm = tag.substr(spos, epos - spos);
|
|
|
253 |
if (attrnm.empty()) {
|
|
|
254 |
m_reason << "Empty attribute name ?? at cpos " << m_pos + epos;
|
|
|
255 |
return false;
|
|
|
256 |
}
|
|
|
257 |
//std::cerr << "attr name [" << attrnm << "]\n";
|
|
|
258 |
skipWS(tag, epos);
|
|
|
259 |
if (epos == std::string::npos || epos == tag.size() - 1 ||
|
|
|
260 |
tag[epos] != '=') {
|
|
|
261 |
m_reason <<"Missing equal sign or value at cpos " << m_pos+epos;
|
|
|
262 |
return false;
|
|
|
263 |
}
|
|
|
264 |
epos++;
|
|
|
265 |
skipWS(tag, epos);
|
|
|
266 |
if (tag[epos] != '"' || epos == tag.size() - 1) {
|
|
|
267 |
m_reason << "Missing dquote or value at cpos " << m_pos+epos;
|
|
|
268 |
return false;
|
|
|
269 |
}
|
|
|
270 |
spos = epos + 1;
|
|
|
271 |
epos = tag.find_first_of("\"", spos);
|
|
|
272 |
if (epos == std::string::npos) {
|
|
|
273 |
m_reason << "Missing closing dquote at cpos " << m_pos+spos;
|
|
|
274 |
return false;
|
|
|
275 |
}
|
|
|
276 |
attrs[attrnm] = tag.substr(spos, epos - spos);
|
|
|
277 |
//std::cerr << "attr value [" << attrs[attrnm] << "]\n";
|
|
|
278 |
if (epos == tag.size() - 1) {
|
|
|
279 |
break;
|
|
|
280 |
}
|
|
|
281 |
epos++;
|
|
|
282 |
skipWS(tag, epos);
|
|
|
283 |
if (epos == tag.size() - 1) {
|
|
|
284 |
break;
|
|
|
285 |
}
|
|
|
286 |
spos = epos;
|
|
|
287 |
}
|
|
|
288 |
tag = tagname;
|
|
|
289 |
return true;
|
|
|
290 |
}
|
|
|
291 |
};
|
|
|
292 |
#endif /* _PICOXML_H_INCLUDED_ */
|