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	if (!url || url[0] == '\0')
 24		return;
 25
 26	ctx.qry.page = NULL;
 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	}
 57}
 58
 59static char *substr(const char *head, const char *tail)
 60{
 61	char *buf;
 62
 63	if (tail < head)
 64		return xstrdup("");
 65	buf = xmalloc(tail - head + 1);
 66	strncpy(buf, head, tail - head);
 67	buf[tail - head] = '\0';
 68	return buf;
 69}
 70
 71static void parse_user(const char *t, char **name, char **email, unsigned long *date, int *tz)
 72{
 73	struct ident_split ident;
 74	unsigned email_len;
 75
 76	if (!split_ident_line(&ident, t, strchrnul(t, '\n') - t)) {
 77		*name = substr(ident.name_begin, ident.name_end);
 78
 79		email_len = ident.mail_end - ident.mail_begin;
 80		*email = xmalloc(strlen("<") + email_len + strlen(">") + 1);
 81		sprintf(*email, "<%.*s>", email_len, ident.mail_begin);
 82
 83		if (ident.date_begin)
 84			*date = strtoul(ident.date_begin, NULL, 10);
 85		if (ident.tz_begin)
 86			*tz = atoi(ident.tz_begin);
 87	}
 88}
 89
 90#ifdef NO_ICONV
 91#define reencode(a, b, c)
 92#else
 93static const char *reencode(char **txt, const char *src_enc, const char *dst_enc)
 94{
 95	char *tmp;
 96
 97	if (!txt)
 98		return NULL;
 99
100	if (!*txt || !src_enc || !dst_enc)
101		return *txt;
102
103	/* no encoding needed if src_enc equals dst_enc */
104	if (!strcasecmp(src_enc, dst_enc))
105		return *txt;
106
107	tmp = reencode_string(*txt, dst_enc, src_enc);
108	if (tmp) {
109		free(*txt);
110		*txt = tmp;
111	}
112	return *txt;
113}
114#endif
115
116static const char *next_header_line(const char *p)
117{
118	p = strchr(p, '\n');
119	if (!p)
120		return NULL;
121	return p + 1;
122}
123
124static int end_of_header(const char *p)
125{
126	return !p || (*p == '\n');
127}
128
129struct commitinfo *cgit_parse_commit(struct commit *commit)
130{
131	const int sha1hex_len = 40;
132	struct commitinfo *ret;
133	const char *p = get_cached_commit_buffer(commit, NULL);
134	const char *t;
135
136	ret = xcalloc(1, sizeof(struct commitinfo));
137	ret->commit = commit;
138
139	if (!p)
140		return ret;
141
142	if (!skip_prefix(p, "tree ", &p))
143		die("Bad commit: %s", oid_to_hex(&commit->object.oid));
144	p += sha1hex_len + 1;
145
146	while (skip_prefix(p, "parent ", &p))
147		p += sha1hex_len + 1;
148
149	if (p && skip_prefix(p, "author ", &p)) {
150		parse_user(p, &ret->author, &ret->author_email,
151			&ret->author_date, &ret->author_tz);
152		p = next_header_line(p);
153	}
154
155	if (p && skip_prefix(p, "committer ", &p)) {
156		parse_user(p, &ret->committer, &ret->committer_email,
157			&ret->committer_date, &ret->committer_tz);
158		p = next_header_line(p);
159	}
160
161	if (p && skip_prefix(p, "encoding ", &p)) {
162		t = strchr(p, '\n');
163		if (t) {
164			ret->msg_encoding = substr(p, t + 1);
165			p = t + 1;
166		}
167	}
168
169	if (!ret->msg_encoding)
170		ret->msg_encoding = xstrdup("UTF-8");
171
172	while (!end_of_header(p))
173		p = next_header_line(p);
174	while (p && *p == '\n')
175		p++;
176	if (!p)
177		return ret;
178
179	t = strchrnul(p, '\n');
180	ret->subject = substr(p, t);
181	while (*t == '\n')
182		t++;
183	ret->msg = xstrdup(t);
184
185	reencode(&ret->author, ret->msg_encoding, PAGE_ENCODING);
186	reencode(&ret->author_email, ret->msg_encoding, PAGE_ENCODING);
187	reencode(&ret->committer, ret->msg_encoding, PAGE_ENCODING);
188	reencode(&ret->committer_email, ret->msg_encoding, PAGE_ENCODING);
189	reencode(&ret->subject, ret->msg_encoding, PAGE_ENCODING);
190	reencode(&ret->msg, ret->msg_encoding, PAGE_ENCODING);
191
192	return ret;
193}
194
195struct taginfo *cgit_parse_tag(struct tag *tag)
196{
197	void *data;
198	enum object_type type;
199	unsigned long size;
200	const char *p;
201	struct taginfo *ret = NULL;
202
203	data = read_object_file(&tag->object.oid, &type, &size);
204	if (!data || type != OBJ_TAG)
205		goto cleanup;
206
207	ret = xcalloc(1, sizeof(struct taginfo));
208
209	for (p = data; !end_of_header(p); p = next_header_line(p)) {
210		if (skip_prefix(p, "tagger ", &p)) {
211			parse_user(p, &ret->tagger, &ret->tagger_email,
212				&ret->tagger_date, &ret->tagger_tz);
213		}
214	}
215
216	while (p && *p == '\n')
217		p++;
218
219	if (p && *p)
220		ret->msg = xstrdup(p);
221
222cleanup:
223	free(data);
224	return ret;
225}