1/* parsing.c: parsing of config files
  2 *
  3 * Copyright (C) 2006-2014 cgit Development Team <cgit@lists.zx2c4.com>
  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 *c, *cmd, *p;
 21	struct cgit_repo *repo;
 22
 23	ctx.repo = NULL;
 24	if (!url || url[0] == '\0')
 25		return;
 26
 27	ctx.repo = cgit_get_repoinfo(url);
 28	if (ctx.repo) {
 29		ctx.qry.repo = ctx.repo->url;
 30		return;
 31	}
 32
 33	cmd = NULL;
 34	c = strchr(url, '/');
 35	while (c) {
 36		c[0] = '\0';
 37		repo = cgit_get_repoinfo(url);
 38		if (repo) {
 39			ctx.repo = repo;
 40			cmd = c;
 41		}
 42		c[0] = '/';
 43		c = strchr(c + 1, '/');
 44	}
 45
 46	if (ctx.repo) {
 47		ctx.qry.repo = ctx.repo->url;
 48		p = strchr(cmd + 1, '/');
 49		if (p) {
 50			p[0] = '\0';
 51			if (p[1])
 52				ctx.qry.path = trim_end(p + 1, '/');
 53		}
 54		if (cmd[1])
 55			ctx.qry.page = xstrdup(cmd + 1);
 56		return;
 57	}
 58}
 59
 60static char *substr(const char *head, const char *tail)
 61{
 62	char *buf;
 63
 64	if (tail < head)
 65		return xstrdup("");
 66	buf = xmalloc(tail - head + 1);
 67	strncpy(buf, head, tail - head);
 68	buf[tail - head] = '\0';
 69	return buf;
 70}
 71
 72static const char *parse_user(const char *t, char **name, char **email, unsigned long *date)
 73{
 74	const char *p = t;
 75	int mode = 1;
 76
 77	while (p && *p) {
 78		if (mode == 1 && *p == '<') {
 79			*name = substr(t, p - 1);
 80			t = p;
 81			mode++;
 82		} else if (mode == 1 && *p == '\n') {
 83			*name = substr(t, p);
 84			p++;
 85			break;
 86		} else if (mode == 2 && *p == '>') {
 87			*email = substr(t, p + 1);
 88			t = p;
 89			mode++;
 90		} else if (mode == 2 && *p == '\n') {
 91			*email = substr(t, p);
 92			p++;
 93			break;
 94		} else if (mode == 3 && isdigit(*p)) {
 95			*date = atol(p);
 96			mode++;
 97		} else if (*p == '\n') {
 98			p++;
 99			break;
100		}
101		p++;
102	}
103	return p;
104}
105
106#ifdef NO_ICONV
107#define reencode(a, b, c)
108#else
109static const char *reencode(char **txt, const char *src_enc, const char *dst_enc)
110{
111	char *tmp;
112
113	if (!txt)
114		return NULL;
115
116	if (!*txt || !src_enc || !dst_enc)
117		return *txt;
118
119	/* no encoding needed if src_enc equals dst_enc */
120	if (!strcasecmp(src_enc, dst_enc))
121		return *txt;
122
123	tmp = reencode_string(*txt, dst_enc, src_enc);
124	if (tmp) {
125		free(*txt);
126		*txt = tmp;
127	}
128	return *txt;
129}
130#endif
131
132struct commitinfo *cgit_parse_commit(struct commit *commit)
133{
134	struct commitinfo *ret;
135	const char *p = get_cached_commit_buffer(commit, NULL);
136	const char *t;
137
138	ret = xmalloc(sizeof(*ret));
139	ret->commit = commit;
140	ret->author = NULL;
141	ret->author_email = NULL;
142	ret->committer = NULL;
143	ret->committer_email = NULL;
144	ret->subject = NULL;
145	ret->msg = NULL;
146	ret->msg_encoding = NULL;
147
148	if (p == NULL)
149		return ret;
150
151	if (!starts_with(p, "tree "))
152		die("Bad commit: %s", sha1_to_hex(commit->object.sha1));
153	else
154		p += 46; // "tree " + hex[40] + "\n"
155
156	while (starts_with(p, "parent "))
157		p += 48; // "parent " + hex[40] + "\n"
158
159	if (p && starts_with(p, "author ")) {
160		p = parse_user(p + 7, &ret->author, &ret->author_email,
161			&ret->author_date);
162	}
163
164	if (p && starts_with(p, "committer ")) {
165		p = parse_user(p + 10, &ret->committer, &ret->committer_email,
166			&ret->committer_date);
167	}
168
169	if (p && starts_with(p, "encoding ")) {
170		p += 9;
171		t = strchr(p, '\n');
172		if (t) {
173			ret->msg_encoding = substr(p, t + 1);
174			p = t + 1;
175		}
176	}
177
178	/* if no special encoding is found, assume UTF-8 */
179	if (!ret->msg_encoding)
180		ret->msg_encoding = xstrdup("UTF-8");
181
182	// skip unknown header fields
183	while (p && *p && (*p != '\n')) {
184		p = strchr(p, '\n');
185		if (p)
186			p++;
187	}
188
189	// skip empty lines between headers and message
190	while (p && *p == '\n')
191		p++;
192
193	if (!p)
194		return ret;
195
196	t = strchr(p, '\n');
197	if (t) {
198		ret->subject = substr(p, t);
199		p = t + 1;
200
201		while (p && *p == '\n') {
202			p = strchr(p, '\n');
203			if (p)
204				p++;
205		}
206		if (p)
207			ret->msg = xstrdup(p);
208	} else
209		ret->subject = xstrdup(p);
210
211	reencode(&ret->author, ret->msg_encoding, PAGE_ENCODING);
212	reencode(&ret->author_email, ret->msg_encoding, PAGE_ENCODING);
213	reencode(&ret->committer, ret->msg_encoding, PAGE_ENCODING);
214	reencode(&ret->committer_email, ret->msg_encoding, PAGE_ENCODING);
215	reencode(&ret->subject, ret->msg_encoding, PAGE_ENCODING);
216	reencode(&ret->msg, ret->msg_encoding, PAGE_ENCODING);
217
218	return ret;
219}
220
221
222struct taginfo *cgit_parse_tag(struct tag *tag)
223{
224	void *data;
225	enum object_type type;
226	unsigned long size;
227	const char *p;
228	struct taginfo *ret;
229
230	data = read_sha1_file(tag->object.sha1, &type, &size);
231	if (!data || type != OBJ_TAG) {
232		free(data);
233		return 0;
234	}
235
236	ret = xmalloc(sizeof(*ret));
237	ret->tagger = NULL;
238	ret->tagger_email = NULL;
239	ret->tagger_date = 0;
240	ret->msg = NULL;
241
242	p = data;
243
244	while (p && *p) {
245		if (*p == '\n')
246			break;
247
248		if (starts_with(p, "tagger ")) {
249			p = parse_user(p + 7, &ret->tagger, &ret->tagger_email,
250				&ret->tagger_date);
251		} else {
252			p = strchr(p, '\n');
253			if (p)
254				p++;
255		}
256	}
257
258	// skip empty lines between headers and message
259	while (p && *p == '\n')
260		p++;
261
262	if (p && *p)
263		ret->msg = xstrdup(p);
264	free(data);
265	return ret;
266}