Annotation of libwww/Library/src/HTParse.c, revision 1.1.1.1
1.1 timbl 1: /* Parse HyperText Document Address HTParse.c
2: ** ================================
3: */
4:
5: #include "HTUtils.h"
6: #include "HTParse.h"
7: #include "tcp.h"
8:
9: struct struct_parts {
10: char * access;
11: char * host;
12: char * absolute;
13: char * relative;
14: /* char * search; no - treated as part of path */
15: char * anchor;
16: };
17:
18:
19: /* Strip white space off a string
20: ** ------------------------------
21: **
22: ** On exit,
23: ** Return value points to first non-white character, or to 0 if none.
24: ** All trailing white space is OVERWRITTEN with zero.
25: */
26:
27: #ifdef __STDC__
28: char * HTStrip(char * s)
29: #else
30: char * HTStrip(s)
31: char *s;
32: #endif
33: {
34: #define SPACE(c) ((c==' ')||(c=='\t')||(c=='\n'))
35: char * p=s;
36: for(p=s;*p;p++); /* Find end of string */
37: for(p--;p>=s;p--) {
38: if(SPACE(*p)) *p=0; /* Zap trailing blanks */
39: else break;
40: }
41: while(SPACE(*s))s++; /* Strip leading blanks */
42: return s;
43: }
44:
45:
46: /* Scan a filename for its consituents
47: ** -----------------------------------
48: **
49: ** On entry,
50: ** name points to a document name which may be incomplete.
51: ** On exit,
52: ** absolute or relative may be nonzero (but not both).
53: ** host, anchor and access may be nonzero if they were specified.
54: ** Any which are nonzero point to zero terminated strings.
55: */
56: #ifdef __STDC__
57: PRIVATE void scan(char * name, struct struct_parts *parts)
58: #else
59: PRIVATE void scan(name, parts)
60: char * name;
61: struct struct_parts *parts;
62: #endif
63: {
64: char * after_access;
65: char * p;
66: int length = strlen(name);
67:
68: parts->access = 0;
69: parts->host = 0;
70: parts->absolute = 0;
71: parts->relative = 0;
72: parts->anchor = 0;
73:
74: after_access = name;
75: for(p=name; *p; p++) {
76: if (*p==':') {
77: *p = 0;
78: parts->access = name; /* Access name has been specified */
79: after_access = p+1;
80: }
81: if (*p=='/') break;
82: if (*p=='#') break;
83: }
84:
85: for(p=name+length-1; p>=name; p--) {
86: if (*p =='#') {
87: parts->anchor=p+1;
88: *p=0; /* terminate the rest */
89: }
90: }
91: p = after_access;
92: if (*p=='/'){
93: if (p[1]=='/') {
94: parts->host = p+2; /* host has been specified */
95: *p=0; /* Terminate access */
96: p=strchr(parts->host,'/'); /* look for end of host name if any */
97: if(p) {
98: *p=0; /* Terminate host */
99: parts->absolute = p+1; /* Root has been found */
100: }
101: } else {
102: parts->absolute = p+1; /* Root found but no host */
103: }
104: } else {
105: parts->relative = (*after_access) ? after_access : 0; /* zero for "" */
106: }
107:
108: /* Access specified but no host: the anchor was not really one
109: e.g. news:j462#36487@foo.bar -- JFG 10/7/92, from bug report */
110: if (parts->access && ! parts->host && parts->anchor) {
111: *(parts->anchor - 1) = '#'; /* Restore the '#' in the address */
112: parts->anchor = 0;
113: }
114:
115: #ifdef NOT_DEFINED /* search is just treated as part of path */
116: {
117: char *p = relative ? relative : absolute;
118: if (p) {
119: char * q = strchr(p, '?'); /* Any search string? */
120: if (q) {
121: *q = 0; /* If so, chop that off. */
122: parts->search = q+1;
123: }
124: }
125: }
126: #endif
127: } /*scan */
128:
129:
130: /* Parse a Name relative to another name
131: ** -------------------------------------
132: **
133: ** This returns those parts of a name which are given (and requested)
134: ** substituting bits from the related name where necessary.
135: **
136: ** On entry,
137: ** aName A filename given
138: ** relatedName A name relative to which aName is to be parsed
139: ** wanted A mask for the bits which are wanted.
140: **
141: ** On exit,
142: ** returns A pointer to a malloc'd string which MUST BE FREED
143: */
144: #ifdef __STDC__
145: char * HTParse(const char * aName, const char * relatedName, int wanted)
146: #else
147: char * HTParse(aName, relatedName, wanted)
148: char * aName;
149: char * relatedName;
150: int wanted;
151: #endif
152:
153: {
154: char * result = 0;
155: char * return_value = 0;
156: int len;
157: char * name = 0;
158: char * rel = 0;
159: char * p;
160: struct struct_parts given, related;
161:
162: /* Make working copies of input strings to cut up:
163: */
164: len = strlen(aName)+strlen(relatedName)+10;
165: result=(char *)malloc(len); /* Lots of space: more than enough */
166: if (result == NULL) outofmem(__FILE__, "HTParse");
167:
168: StrAllocCopy(name, aName);
169: StrAllocCopy(rel, relatedName);
170:
171: scan(name, &given);
172: scan(rel, &related);
173: result[0]=0; /* Clear string */
174: if (wanted & PARSE_ACCESS)
175: if (given.access|| related.access) {
176: strcat(result, given.access ? given.access : related.access);
177: if(wanted & PARSE_PUNCTUATION) strcat(result, ":");
178: }
179:
180: if (given.access && related.access) /* If different, inherit nothing. */
181: if (strcmp(given.access, related.access)!=0) {
182: related.host=0;
183: related.absolute=0;
184: related.relative=0;
185: related.anchor=0;
186: }
187:
188: if (wanted & PARSE_HOST)
189: if(given.host || related.host) {
190: if(wanted & PARSE_PUNCTUATION) strcat(result, "//");
191: strcat(result, given.host ? given.host : related.host);
192: }
193:
194: if (given.host && related.host) /* If different hosts, inherit no path. */
195: if (strcmp(given.host, related.host)!=0) {
196: related.absolute=0;
197: related.relative=0;
198: related.anchor=0;
199: }
200:
201: if (wanted & PARSE_PATH) {
202: if(given.absolute) { /* All is given */
203: if(wanted & PARSE_PUNCTUATION) strcat(result, "/");
204: strcat(result, given.absolute);
205: } else if(related.absolute) { /* Adopt path not name */
206: strcat(result, "/");
207: strcat(result, related.absolute);
208: if (given.relative) {
209: p = strchr(result, '?'); /* Search part? */
210: if (!p) p=result+strlen(result)-1;
211: for (; *p!='/'; p--); /* last / */
212: p[1]=0; /* Remove filename */
213: strcat(result, given.relative); /* Add given one */
214: HTSimplify (result);
215: }
216: } else if(given.relative) {
217: strcat(result, given.relative); /* what we've got */
218: } else if(related.relative) {
219: strcat(result, related.relative);
220: } else { /* No inheritance */
221: strcat(result, "/");
222: }
223: }
224:
225: if (wanted & PARSE_ANCHOR)
226: if(given.anchor || related.anchor) {
227: if(wanted & PARSE_PUNCTUATION) strcat(result, "#");
228: strcat(result, given.anchor ? given.anchor : related.anchor);
229: }
230: free(rel);
231: free(name);
232:
233: StrAllocCopy(return_value, result);
234: free(result);
235: return return_value; /* exactly the right length */
236: }
237:
238: /* Simplify a filename
239: // -------------------
240: //
241: // A unix-style file is allowed to contain the seqeunce xxx/../ which may be
242: // replaced by "" , and the seqeunce "/./" which may be replaced by "/".
243: // Simplification helps us recognize duplicate filenames.
244: //
245: // Thus, /etc/junk/../fred becomes /etc/fred
246: // /etc/junk/./fred becomes /etc/junk/fred
247: */
248: #ifdef __STDC__
249: void HTSimplify(char * filename)
250: #else
251: void HTSimplify(filename)
252: char * filename;
253: #endif
254:
255: {
256: char * p;
257: char * q;
258: for(p=filename+2; *p; p++) {
259: if (*p=='/') {
260: if ((p[1]=='.') && (p[2]=='.') && (p[3]=='/' || !p[3] )) {
261: for (q=p-1; (q>filename) && (*q!='/'); q--); /* prev slash */
262: if (*q=='/') {
263: strcpy(q, p+3); /* Remove /xxx/.. */
264: if (!*filename) strcpy(filename, "/");
265: p = q-1; /* Start again with prev slash */
266: } else { /* xxx/.. error? */
267: strcpy(filename, p[3] ? p+4 : p+3); /* rm xxx/../ */
268: p = filename; /* Start again */
269: }
270: } else if ((p[1]=='.') && (p[2]=='/' || !p[2])) {
271: strcpy(p, p+2); /* Remove a slash and a dot */
272: }
273: }
274: }
275: }
276:
277:
278: /* Make Relative Name
279: ** ------------------
280: **
281: ** This function creates and returns a string which gives an expression of
282: ** one address as related to another. Where there is no relation, an absolute
283: ** address is retured.
284: **
285: ** On entry,
286: ** Both names must be absolute, fully qualified names of nodes
287: ** (no anchor bits)
288: **
289: ** On exit,
290: ** The return result points to a newly allocated name which, if
291: ** parsed by HTParse relative to relatedName, will yield aName.
292: ** The caller is responsible for freeing the resulting name later.
293: **
294: */
295: #ifdef __STDC__
296: char * HTRelative(const char * aName, const char *relatedName)
297: #else
298: char * HTRelative(aName, relatedName)
299: char * aName;
300: char * relatedName;
301: #endif
302: {
303: char * result = 0;
304: CONST char *p = aName;
305: CONST char *q = relatedName;
306: CONST char * after_access = 0;
307: CONST char * path = 0;
308: CONST char * last_slash = 0;
309: int slashes = 0;
310:
311: for(;*p; p++, q++) { /* Find extent of match */
312: if (*p!=*q) break;
313: if (*p==':') after_access = p+1;
314: if (*p=='/') {
315: last_slash = p;
316: slashes++;
317: if (slashes==3) path=p;
318: }
319: }
320:
321: /* q, p point to the first non-matching character or zero */
322:
323: if (!after_access) { /* Different access */
324: StrAllocCopy(result, aName);
325: } else if (slashes<3){ /* Different nodes */
326: StrAllocCopy(result, after_access);
327: } else if (slashes==3){ /* Same node, different path */
328: StrAllocCopy(result, path);
329: } else { /* Some path in common */
330: int levels= 0;
331: for(; *q && (*q!='#'); q++) if (*q=='/') levels++;
332: result = (char *)malloc(3*levels + strlen(last_slash) + 1);
333: if (result == NULL) outofmem(__FILE__, "HTRelative");
334: result[0]=0;
335: for(;levels; levels--)strcat(result, "../");
336: strcat(result, last_slash+1);
337: }
338: if (TRACE) fprintf(stderr, "HT: `%s' expressed relative to\n `%s' is\n `%s'.",
339: aName, relatedName, result);
340: return result;
341: }
Webmaster