1/* config.c: parsing of config files
  2 *
  3 * Copyright (C) 2006 Lars Hjemli
  4 *
  5 * Licensed under GNU General Public License v2
  6 *   (see COPYING for full license text)
  7 */
  8
  9#include "cgit.h"
 10
 11/*
 12 * url syntax: [repo ['/' cmd [ '/' path]]]
 13 *   repo: any valid repo url, may contain '/'
 14 *   cmd:  log | commit | diff | tree | view | blob | snapshot
 15 *   path: any valid path, may contain '/'
 16 *
 17 */
 18void cgit_parse_url(const char *url)
 19{
 20	char *cmd, *p;
 21
 22	ctx.repo = NULL;
 23	if (!url || url[0] == '\0')
 24		return;
 25
 26	ctx.repo = cgit_get_repoinfo(url);
 27	if (ctx.repo) {
 28		ctx.qry.repo = ctx.repo->url;
 29		return;
 30	}
 31
 32	cmd = strchr(url, '/');
 33	while (!ctx.repo && cmd) {
 34		cmd[0] = '\0';
 35		ctx.repo = cgit_get_repoinfo(url);
 36		if (ctx.repo == NULL) {
 37			cmd[0] = '/';
 38			cmd = strchr(cmd + 1, '/');
 39			continue;
 40		}
 41
 42		ctx.qry.repo = ctx.repo->url;
 43		p = strchr(cmd + 1, '/');
 44		if (p) {
 45			p[0] = '\0';
 46			if (p[1])
 47				ctx.qry.path = trim_end(p + 1, '/');
 48		}
 49		if (cmd[1])
 50			ctx.qry.page = xstrdup(cmd + 1);
 51		return;
 52	}
 53}
 54
 55char *substr(const char *head, const char *tail)
 56{
 57	char *buf;
 58
 59	buf = xmalloc(tail - head + 1);
 60	strncpy(buf, head, tail - head);
 61	buf[tail - head] = '\0';
 62	return buf;
 63}
 64
 65char *parse_user(char *t, char **name, char **email, unsigned long *date)
 66{
 67	char *p = t;
 68	int mode = 1;
 69
 70	while (p && *p) {
 71		if (mode == 1 && *p == '<') {
 72			*name = substr(t, p - 1);
 73			t = p;
 74			mode++;
 75		} else if (mode == 1 && *p == '\n') {
 76			*name = substr(t, p);
 77			p++;
 78			break;
 79		} else if (mode == 2 && *p == '>') {
 80			*email = substr(t, p + 1);
 81			t = p;
 82			mode++;
 83		} else if (mode == 2 && *p == '\n') {
 84			*email = substr(t, p);
 85			p++;
 86			break;
 87		} else if (mode == 3 && isdigit(*p)) {
 88			*date = atol(p);
 89			mode++;
 90		} else if (*p == '\n') {
 91			p++;
 92			break;
 93		}
 94		p++;
 95	}
 96	return p;
 97}
 98
 99#ifdef NO_ICONV
100#define reencode(a, b, c)
101#else
102const char *reencode(char **txt, const char *src_enc, const char *dst_enc)
103{
104	char *tmp;
105
106	if (!txt || !*txt || !src_enc || !dst_enc)
107		return *txt;
108
109	/* no encoding needed if src_enc equals dst_enc */
110	if(!strcasecmp(src_enc, dst_enc))
111		return *txt;
112
113	tmp = reencode_string(*txt, dst_enc, src_enc);
114	if (tmp) {
115		free(*txt);
116		*txt = tmp;
117	}
118	return *txt;
119}
120#endif
121
122struct commitinfo *cgit_parse_commit(struct commit *commit)
123{
124	struct commitinfo *ret;
125	char *p = commit->buffer, *t = commit->buffer;
126
127	ret = xmalloc(sizeof(*ret));
128	ret->commit = commit;
129	ret->author = NULL;
130	ret->author_email = NULL;
131	ret->committer = NULL;
132	ret->committer_email = NULL;
133	ret->subject = NULL;
134	ret->msg = NULL;
135	ret->msg_encoding = NULL;
136
137	if (p == NULL)
138		return ret;
139
140	if (strncmp(p, "tree ", 5))
141		die("Bad commit: %s", sha1_to_hex(commit->object.sha1));
142	else
143		p += 46; // "tree " + hex[40] + "\n"
144
145	while (!strncmp(p, "parent ", 7))
146		p += 48; // "parent " + hex[40] + "\n"
147
148	if (p && !strncmp(p, "author ", 7)) {
149		p = parse_user(p + 7, &ret->author, &ret->author_email,
150			&ret->author_date);
151	}
152
153	if (p && !strncmp(p, "committer ", 9)) {
154		p = parse_user(p + 9, &ret->committer, &ret->committer_email,
155			&ret->committer_date);
156	}
157
158	if (p && !strncmp(p, "encoding ", 9)) {
159		p += 9;
160		t = strchr(p, '\n');
161		if (t) {
162			ret->msg_encoding = substr(p, t + 1);
163			p = t + 1;
164		}
165	}
166
167	/* if no special encoding is found, assume UTF-8 */
168	if(!ret->msg_encoding)
169		ret->msg_encoding = xstrdup("UTF-8");
170
171	// skip unknown header fields
172	while (p && *p && (*p != '\n')) {
173		p = strchr(p, '\n');
174		if (p)
175			p++;
176	}
177
178	// skip empty lines between headers and message
179	while (p && *p == '\n')
180		p++;
181
182	if (!p)
183		return ret;
184
185	t = strchr(p, '\n');
186	if (t) {
187		ret->subject = substr(p, t);
188		p = t + 1;
189
190		while (p && *p == '\n') {
191			p = strchr(p, '\n');
192			if (p)
193				p++;
194		}
195		if (p)
196			ret->msg = xstrdup(p);
197	} else
198		ret->subject = xstrdup(p);
199
200	reencode(&ret->author, ret->msg_encoding, PAGE_ENCODING);
201	reencode(&ret->author_email, ret->msg_encoding, PAGE_ENCODING);
202	reencode(&ret->committer, ret->msg_encoding, PAGE_ENCODING);
203	reencode(&ret->committer_email, ret->msg_encoding, PAGE_ENCODING);
204	reencode(&ret->subject, ret->msg_encoding, PAGE_ENCODING);
205	reencode(&ret->msg, ret->msg_encoding, PAGE_ENCODING);
206
207	return ret;
208}
209
210
211struct taginfo *cgit_parse_tag(struct tag *tag)
212{
213	void *data;
214	enum object_type type;
215	unsigned long size;
216	char *p;
217	struct taginfo *ret;
218
219	data = read_sha1_file(tag->object.sha1, &type, &size);
220	if (!data || type != OBJ_TAG) {
221		free(data);
222		return 0;
223	}
224
225	ret = xmalloc(sizeof(*ret));
226	ret->tagger = NULL;
227	ret->tagger_email = NULL;
228	ret->tagger_date = 0;
229	ret->msg = NULL;
230
231	p = data;
232
233	while (p && *p) {
234		if (*p == '\n')
235			break;
236
237		if (!strncmp(p, "tagger ", 7)) {
238			p = parse_user(p + 7, &ret->tagger, &ret->tagger_email,
239				&ret->tagger_date);
240		} else {
241			p = strchr(p, '\n');
242			if (p)
243				p++;
244		}
245	}
246
247	// skip empty lines between headers and message
248	while (p && *p == '\n')
249		p++;
250
251	if (p && *p)
252		ret->msg = xstrdup(p);
253	free(data);
254	return ret;
255}