Routino SVN Repository Browser

Check out the latest version of Routino: svn co http://routino.org/svn/trunk routino

/[routino]/trunk/src/xmlparse.l
ViewVC logotype

Annotation of /trunk/src/xmlparse.l

Parent Directory Parent Directory | Revision Log Revision Log

Revision 344 - (hide annotations) (download)
Wed Mar 31 17:19:22 2010 UTC (14 years, 11 months ago) by amb
File size: 11597 byte(s) 
Call the XML tag functions for the end tags as well as the start tags.
1 amb 334 %{
2     /***************************************
3 amb 344 $Header: /home/amb/CVS/routino/src/xmlparse.l,v 1.3 2010-03-31 17:18:27 amb Exp $
4 amb 334
5     A simple generic XML parser where the structure comes from the function parameters.
6    
7     Part of the Routino routing software.
8     ******************/ /******************
9     This file Copyright 2010 Andrew M. Bishop
10    
11     This program is free software: you can redistribute it and/or modify
12     it under the terms of the GNU Affero General Public License as published by
13     the Free Software Foundation, either version 3 of the License, or
14     (at your option) any later version.
15    
16     This program is distributed in the hope that it will be useful,
17     but WITHOUT ANY WARRANTY; without even the implied warranty of
18     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
19     GNU Affero General Public License for more details.
20    
21     You should have received a copy of the GNU Affero General Public License
22     along with this program. If not, see <http://www.gnu.org/licenses/>.
23     ***************************************/
24    
25    
26     #include <stdio.h>
27     #include <stdlib.h>
28     #include <string.h>
29    
30     /* Parser outputs */
31    
32     #define LEX_TAG_BEGIN 1
33     #define LEX_TAG_POP 2
34     #define LEX_TAG_PUSH 3
35     #define LEX_TAG_FINISH 4
36     #define LEX_ATTR_KEY 5
37     #define LEX_ATTR_VAL 6
38    
39    
40     /* Lexer definitions */
41    
42     #define YY_SKIP_YYWRAP 1 /* Remove error with prototype of ..._yywrap */
43     #ifndef yywrap
44     /*+ Needed in lex but does nothing. +*/
45     #define yywrap() 1
46     #endif
47    
48     /*+ Reset the current string. +*/
49     #define reset_string \
50     if(string) *string=0; \
51     stringused=0;
52    
53     /*+ append information to the current string. +*/
54     #define append_string(xx) \
55     newlen=strlen(xx); \
56     if((stringused+newlen)>=stringlen) \
57     string=(char*)realloc((void*)string,stringlen=(stringused+newlen+16)); \
58     strcpy(string+stringused,xx); \
59     stringused+=newlen;
60    
61     #define YY_NO_INPUT
62    
63    
64     /* Lexer functions and variables */
65    
66     extern int yylex(void);
67    
68     static char *yylval=NULL;
69    
70     %}
71    
72 amb 344 %option 8bit
73     %option batch
74     %option yylineno
75     %option nodefault
76     %option perf-report
77     %option fast
78    
79 amb 334 W [ \t]
80    
81     nonascii [\200-\377]
82     ascii [ -~]
83     alphanum [a-zA-Z0-9]
84     punct [][!\"#$%&\'()*+,-./:;<=>?@\\^_`{|}~]
85     safepunct [][!\#$%\()*+,-./:;=?@\\^_`{|}~]
86    
87     tag ({alphanum}|[-:])+
88     key ({alphanum}|[-:])+
89     val ({alphanum}|{nonascii}|{safepunct})+
90    
91     %x COMMENT
92     %x TAG_START TAG TAG_ATTR_KEY TAG_ATTR_VAL
93     %x DQUOTED SQUOTED
94    
95     %%
96     /* Must use static variables since the parser returns often. */
97     static char *string=NULL;
98     static int stringlen=0,stringused=0;
99     int newlen;
100    
101     /* Handle comments and other tags */
102    
103     "<!--" { BEGIN(COMMENT); }
104     "<" { BEGIN(TAG_START); }
105 amb 344 [^<]+ { }
106 amb 334
107     /* Comments - not strictly correct. */
108    
109     <COMMENT>"--"{W}*">" { BEGIN(INITIAL); }
110     <COMMENT>">" { }
111     <COMMENT>"-" { }
112 amb 344 <COMMENT>[^->]+ { }
113 amb 334
114     /* Tags */
115    
116     <TAG_START>{W}+ { }
117     <TAG_START>"?xml" { BEGIN(TAG); yylval=yytext; return(LEX_TAG_BEGIN); }
118     <TAG_START>{tag} { BEGIN(TAG); yylval=yytext; return(LEX_TAG_BEGIN); }
119    
120     <TAG_START>"/"{tag}">" { BEGIN(INITIAL); return(LEX_TAG_POP); }
121    
122 amb 344 <TAG_START>.|\n { BEGIN(INITIAL); }
123 amb 334
124     <TAG>{W}+ { }
125     <TAG>"/>" { BEGIN(INITIAL); return(LEX_TAG_FINISH); }
126     <TAG>"?>" { BEGIN(INITIAL); return(LEX_TAG_FINISH); }
127     <TAG>">" { BEGIN(INITIAL); return(LEX_TAG_PUSH); }
128     <TAG>{key} { BEGIN(TAG_ATTR_KEY); yylval=yytext; return(LEX_ATTR_KEY); }
129 amb 344 <TAG>.|\n { }
130 amb 334
131     <TAG_ATTR_KEY>{W}*= { BEGIN(TAG_ATTR_VAL); }
132 amb 344 <TAG_ATTR_KEY>.|\n { BEGIN(TAG); unput(yytext[0]); yylval=NULL; return(LEX_ATTR_VAL); }
133 amb 334
134     <TAG_ATTR_VAL>{W}+ { }
135     <TAG_ATTR_VAL>\" { BEGIN(DQUOTED); reset_string; }
136     <TAG_ATTR_VAL>\' { BEGIN(SQUOTED); reset_string; }
137     <TAG_ATTR_VAL>{val} { BEGIN(TAG); yylval=yytext; return(LEX_ATTR_VAL); }
138 amb 344 <TAG_ATTR_VAL>.|\n { BEGIN(TAG); unput(yytext[0]); yylval=NULL; return(LEX_ATTR_VAL); }
139 amb 334
140     /* Quoted strings */
141    
142     <DQUOTED>\\\\ { append_string(yytext); }
143     <DQUOTED>\\\" { append_string(yytext); }
144     <DQUOTED>\\ { append_string(yytext); }
145     <DQUOTED>\" { BEGIN(TAG); yylval=string; return(LEX_ATTR_VAL); }
146 amb 344 <DQUOTED>[^\\\"]+ { append_string(yytext); }
147 amb 334
148     <SQUOTED>\\\\ { append_string(yytext); }
149     <SQUOTED>\\\' { append_string(yytext); }
150     <SQUOTED>\\ { append_string(yytext); }
151     <SQUOTED>\' { BEGIN(TAG); yylval=string; return(LEX_ATTR_VAL); }
152 amb 344 <SQUOTED>[^\\\']+ { append_string(yytext); }
153 amb 334
154     /* End of file */
155    
156     <<EOF>> { free(string); string=NULL; BEGIN(INITIAL); return(0); }
157    
158     %%
159    
160     #include "xmlparse.h"
161    
162 amb 344 static inline void call_callback(char *name,void (*callback)(),int type,int nattributes,char *attributes[XMLPARSE_MAX_ATTRS])
163     {
164     switch(nattributes)
165     {
166     case 0: (*callback)(type); break;
167     case 1: (*callback)(type,attributes[0]); break;
168     case 2: (*callback)(type,attributes[0],attributes[1]); break;
169     case 3: (*callback)(type,attributes[0],attributes[1],attributes[2]); break;
170     case 4: (*callback)(type,attributes[0],attributes[1],attributes[2],attributes[3]); break;
171     case 5: (*callback)(type,attributes[0],attributes[1],attributes[2],attributes[3],attributes[4]); break;
172     case 6: (*callback)(type,attributes[0],attributes[1],attributes[2],attributes[3],attributes[4],attributes[5]); break;
173     case 7: (*callback)(type,attributes[0],attributes[1],attributes[2],attributes[3],attributes[4],attributes[5],attributes[6]); break;
174     case 8: (*callback)(type,attributes[0],attributes[1],attributes[2],attributes[3],attributes[4],attributes[5],attributes[6],attributes[7]); break;
175     case 9: (*callback)(type,attributes[0],attributes[1],attributes[2],attributes[3],attributes[4],attributes[5],attributes[6],attributes[7],attributes[8]); break;
176     case 10: (*callback)(type,attributes[0],attributes[1],attributes[2],attributes[3],attributes[4],attributes[5],attributes[6],attributes[7],attributes[8],attributes[9]); break;
177     case 11: (*callback)(type,attributes[0],attributes[1],attributes[2],attributes[3],attributes[4],attributes[5],attributes[6],attributes[7],attributes[8],attributes[9],attributes[10]); break;
178     case 12: (*callback)(type,attributes[0],attributes[1],attributes[2],attributes[3],attributes[4],attributes[5],attributes[6],attributes[7],attributes[8],attributes[9],attributes[10],attributes[11]); break;
179     case 13: (*callback)(type,attributes[0],attributes[1],attributes[2],attributes[3],attributes[4],attributes[5],attributes[6],attributes[7],attributes[8],attributes[9],attributes[10],attributes[11],attributes[12]); break;
180     case 14: (*callback)(type,attributes[0],attributes[1],attributes[2],attributes[3],attributes[4],attributes[5],attributes[6],attributes[7],attributes[8],attributes[9],attributes[10],attributes[11],attributes[12],attributes[13]); break;
181     case 15: (*callback)(type,attributes[0],attributes[1],attributes[2],attributes[3],attributes[4],attributes[5],attributes[6],attributes[7],attributes[8],attributes[9],attributes[10],attributes[11],attributes[12],attributes[13],attributes[14]); break;
182     case 16: (*callback)(type,attributes[0],attributes[1],attributes[2],attributes[3],attributes[4],attributes[5],attributes[6],attributes[7],attributes[8],attributes[9],attributes[10],attributes[11],attributes[12],attributes[13],attributes[14],attributes[15]); break;
183 amb 334
184 amb 344 default:
185     fprintf(stderr,"XML Parser: Error too many attributes for tag '%s' on line %d.\n",name,yylineno);
186     exit(1);
187     }
188     }
189    
190    
191 amb 334 /*++++++++++++++++++++++++++++++++++++++
192     Parse the XML and call the functions for each tag as seen.
193    
194     FILE *file The file to parse.
195    
196     xmltag **tags The array of pointers to tags for the top level.
197 amb 337
198     int ignore_unknown_attributes If set to 0 then exit if unknown attribute is seen, if sete to 1 then warn, if set to 2 then ignore.
199 amb 334 ++++++++++++++++++++++++++++++++++++++*/
200    
201 amb 337 void ParseXML(FILE *file,xmltag **tags,int ignore_unknown_attributes)
202 amb 334 {
203     int yychar,i;
204    
205     int nattributes=0;
206     char *attributes[XMLPARSE_MAX_ATTRS];
207     int attribute=0;
208    
209     int stackdepth=0,stackused=0;
210 amb 344 xmltag ***tags_stack=NULL;
211     xmltag **tag_stack=NULL;
212 amb 334 xmltag *tag=NULL;
213    
214     static int first=1;
215    
216     /* Parser (re)-initialisation */
217    
218     yyin=file;
219    
220     if(!first)
221     yyrestart(NULL);
222    
223     first=0;
224    
225     /* The actual parser. */
226    
227     while((yychar=yylex()))
228     switch(yychar)
229     {
230     /* The start of a tag for an element */
231    
232     case LEX_TAG_BEGIN:
233    
234     tag=NULL;
235    
236     for(i=0;tags[i];i++)
237     if(!strcasecmp(yylval,tags[i]->name))
238     {
239     tag=tags[i];
240    
241     for(i=0;i<nattributes;i++)
242     free(attributes[i]);
243    
244     for(i=0;i<XMLPARSE_MAX_ATTRS;i++)
245     if(!tag->attributes[i])
246     break;
247    
248     nattributes=i;
249    
250     for(i=0;i<nattributes;i++)
251     attributes[i]=NULL;
252    
253     break;
254     }
255    
256     if(tag==NULL)
257     {
258 amb 344 fprintf(stderr,"XML Parser: Error unexpected tag '%s' on line %d.\n",yylval,yylineno);
259 amb 334 exit(1);
260     }
261     break;
262    
263     /* The end of the start-tag for an element */
264    
265     case LEX_TAG_PUSH:
266    
267     if(stackused==stackdepth)
268 amb 344 {
269     tag_stack =(xmltag**) realloc((void*)tag_stack ,(stackdepth+=8)*sizeof(xmltag*));
270     tags_stack=(xmltag***)realloc((void*)tags_stack,(stackdepth+=8)*sizeof(xmltag**));
271     }
272 amb 334
273 amb 344 tag_stack [stackused]=tag;
274     tags_stack[stackused]=tags;
275     stackused++;
276    
277     if(tag->callback)
278     call_callback(tag->name,tag->callback,XMLPARSE_TAG_START,nattributes,attributes);
279    
280 amb 334 tags=tag->subtags;
281    
282 amb 344 break;
283    
284 amb 334 /* The end of the empty-element-tag for an element */
285    
286     case LEX_TAG_FINISH:
287    
288     if(tag->callback)
289 amb 344 call_callback(tag->name,tag->callback,XMLPARSE_TAG_START|XMLPARSE_TAG_END,nattributes,attributes);
290 amb 334
291 amb 344 if(stackused>0)
292     tag=tag_stack[stackused-1];
293     else
294     tag=NULL;
295 amb 334
296     break;
297    
298     /* The end of the end-tag for an element */
299    
300     case LEX_TAG_POP:
301    
302 amb 344 stackused--;
303     tags=tags_stack[stackused];
304     tag =tag_stack [stackused];
305 amb 334
306 amb 344 if(tag->callback)
307     call_callback(tag->name,tag->callback,XMLPARSE_TAG_END,nattributes,attributes);
308    
309     tag=tag_stack[stackused-1];
310    
311 amb 334 break;
312    
313     /* An attribute key */
314    
315     case LEX_ATTR_KEY:
316    
317     attribute=-1;
318    
319     for(i=0;i<nattributes;i++)
320     if(!strcasecmp(yylval,tag->attributes[i]))
321     {
322     attribute=i;
323    
324     break;
325     }
326    
327     if(attribute==-1)
328     {
329 amb 337 if(ignore_unknown_attributes==0)
330     {
331 amb 344 fprintf(stderr,"XML Parser: Error unexpected attribute '%s' for tag '%s' on line %d.\n",yylval,tag->name,yylineno);
332 amb 337 exit(1);
333     }
334     else if(ignore_unknown_attributes==1)
335 amb 344 fprintf(stderr,"XML Parser: Warning unexpected attribute '%s' for tag '%s' on line %d.\n",yylval,tag->name,yylineno);
336 amb 334 }
337     break;
338    
339     /* An attribute value */
340    
341     case LEX_ATTR_VAL:
342    
343 amb 337 if(yylval && attribute!=-1)
344 amb 334 attributes[attribute]=strcpy(malloc(strlen(yylval)+1),yylval);
345     }
346    
347     /* Delete the tagdata */
348    
349     for(i=0;i<nattributes;i++)
350     if(attributes[i])
351     free(attributes[i]);
352    
353     if(stackdepth)
354 amb 344 free(tags_stack);
355 amb 334 }

Properties

Name Value
cvs:description A simple generic XML parser.
Index: This Site Routino Software Download Source Code Documentation XML Files Links
Index: View SVN Development Recent changes Development (destination-access branch) Version 3.4.1 (destination-access branch) Version 3.4.1 Version 3.4 (destination-access branch) Version 3.4 Version 3.3.3 (destination-access branch) Version 3.3.3 Version 3.3.2 (destination-access branch) Version 3.3.2 Version 3.3.1 (destination-access branch) Version 3.3.1 Version 3.3 (destination-access branch) Version 3.3 Version 3.2 Version 3.1.1 Version 3.1 Version 3.0 Version 2.7.3 Version 2.7.2 Version 2.7.1 Version 2.7 Version 2.6 Version 2.5.1 Version 2.5 Version 2.4.1 Version 2.4 Version 2.3.2 Version 2.3.1 Version 2.3 Version 2.2 Version 2.1.2 Version 2.1.1 Version 2.1 Version 2.0.3 Version 2.0.2 Version 2.0.1 Version 2.0 Version 1.5.1 Version 1.5 Version 1.4.1 Version 1.4 Version 1.3 Version 1.2 Version 1.1 Version 1.0
Powered by ViewVC 1.3.0-dev ViewVC Help
© Andrew M. Bishop = < amb @ ... >
Root