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	ctx.qry.page = NULL;
 25	if (!url || url[0] == '\0')
 26		return;
 27
 28	ctx.repo = cgit_get_repoinfo(url);
 29	if (ctx.repo) {
 30		ctx.qry.repo = ctx.repo->url;
 31		return;
 32	}
 33
 34	cmd = NULL;
 35	c = strchr(url, '/');
 36	while (c) {
 37		c[0] = '\0';
 38		repo = cgit_get_repoinfo(url);
 39		if (repo) {
 40			ctx.repo = repo;
 41			cmd = c;
 42		}
 43		c[0] = '/';
 44		c = strchr(c + 1, '/');
 45	}
 46
 47	if (ctx.repo) {
 48		ctx.qry.repo = ctx.repo->url;
 49		p = strchr(cmd + 1, '/');
 50		if (p) {
 51			p[0] = '\0';
 52			if (p[1])
 53				ctx.qry.path = trim_end(p + 1, '/');
 54		}
 55		if (cmd[1])
 56			ctx.qry.page = xstrdup(cmd + 1);
 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 void parse_user(const char *t, char **name, char **email, unsigned long *date, int *tz)
 73{
 74	struct ident_split ident;
 75	unsigned email_len;
 76
 77	if (!split_ident_line(&ident, t, strchrnul(t, '\n') - t)) {
 78		*name = substr(ident.name_begin, ident.name_end);
 79
 80		email_len = ident.mail_end - ident.mail_begin;
 81		*email = xmalloc(strlen("<") + email_len + strlen(">") + 1);
 82		sprintf(*email, "<%.*s>", email_len, ident.mail_begin);
 83
 84		if (ident.date_begin)
 85			*date = strtoul(ident.date_begin, NULL, 10);
 86		if (ident.tz_begin)
 87			*tz = atoi(ident.tz_begin);
 88	}
 89}
 90
 91#ifdef NO_ICONV
 92#define reencode(a, b, c)
 93#else
 94static const char *reencode(char **txt, const char *src_enc, const char *dst_enc)
 95{
 96	char *tmp;
 97
 98	if (!txt)
 99		return NULL;
100
101	if (!*txt || !src_enc || !dst_enc)
102		return *txt;
103
104	/* no encoding needed if src_enc equals dst_enc */
105	if (!strcasecmp(src_enc, dst_enc))
106		return *txt;
107
108	tmp = reencode_string(*txt, dst_enc, src_enc);
109	if (tmp) {
110		free(*txt);
111		*txt = tmp;
112	}
113	return *txt;
114}
115#endif
116
117static const char *next_header_line(const char *p)
118{
119	p = strchr(p, '\n');
120	if (!p)
121		return NULL;
122	return p + 1;
123}
124
125static int end_of_header(const char *p)
126{
127	return !p || (*p == '\n');
128}
129
130struct commitinfo *cgit_parse_commit(struct commit *commit)
131{
132	const int sha1hex_len = 40;
133	struct commitinfo *ret;
134	const char *p = get_cached_commit_buffer(commit, NULL);
135	const char *t;
136
137	ret = xcalloc(1, sizeof(struct commitinfo));
138	ret->commit = commit;
139
140	if (!p)
141		return ret;
142
143	if (!skip_prefix(p, "tree ", &p))
144		die("Bad commit: %s", oid_to_hex(&commit->object.oid));
145	p += sha1hex_len + 1;
146
147	while (skip_prefix(p, "parent ", &p))
148		p += sha1hex_len + 1;
149
150	if (p && skip_prefix(p, "author ", &p)) {
151		parse_user(p, &ret->author, &ret->author_email,
152			&ret->author_date, &ret->author_tz);
153		p = next_header_line(p);
154	}
155
156	if (p && skip_prefix(p, "committer ", &p)) {
157		parse_user(p, &ret->committer, &ret->committer_email,
158			&ret->committer_date, &ret->committer_tz);
159		p = next_header_line(p);
160	}
161
162	if (p && skip_prefix(p, "encoding ", &p)) {
163		t = strchr(p, '\n');
164		if (t) {
165			ret->msg_encoding = substr(p, t + 1);
166			p = t + 1;
167		}
168	}
169
170	if (!ret->msg_encoding)
171		ret->msg_encoding = xstrdup("UTF-8");
172
173	while (!end_of_header(p))
174		p = next_header_line(p);
175	while (p && *p == '\n')
176		p++;
177	if (!p)
178		return ret;
179
180	t = strchrnul(p, '\n');
181	ret->subject = substr(p, t);
182	while (*t == '\n')
183		t++;
184	ret->msg = xstrdup(t);
185
186	reencode(&ret->author, ret->msg_encoding, PAGE_ENCODING);
187	reencode(&ret->author_email, ret->msg_encoding, PAGE_ENCODING);
188	reencode(&ret->committer, ret->msg_encoding, PAGE_ENCODING);
189	reencode(&ret->committer_email, ret->msg_encoding, PAGE_ENCODING);
190	reencode(&ret->subject, ret->msg_encoding, PAGE_ENCODING);
191	reencode(&ret->msg, ret->msg_encoding, PAGE_ENCODING);
192
193	return ret;
194}
195
196struct taginfo *cgit_parse_tag(struct tag *tag)
197{
198	void *data;
199	enum object_type type;
200	unsigned long size;
201	const char *p;
202	struct taginfo *ret = NULL;
203
204	data = read_sha1_file(tag->object.oid.hash, &type, &size);
205	if (!data || type != OBJ_TAG)
206		goto cleanup;
207
208	ret = xcalloc(1, sizeof(struct taginfo));
209
210	for (p = data; !end_of_header(p); p = next_header_line(p)) {
211		if (skip_prefix(p, "tagger ", &p)) {
212			parse_user(p, &ret->tagger, &ret->tagger_email,
213				&ret->tagger_date, &ret->tagger_tz);
214		}
215	}
216
217	while (p && *p == '\n')
218		p++;
219
220	if (p && *p)
221		ret->msg = xstrdup(p);
222
223cleanup:
224	free(data);
225	return ret;
226}