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 *line_end = strchrnul(t, '\n');
 75	struct ident_split ident;
 76	unsigned email_len;
 77
 78	if (!split_ident_line(&ident, t, line_end - t)) {
 79		*name = substr(ident.name_begin, ident.name_end);
 80
 81		email_len = ident.mail_end - ident.mail_begin;
 82		*email = xmalloc(strlen("<") + email_len + strlen(">") + 1);
 83		sprintf(*email, "<%.*s>", email_len, ident.mail_begin);
 84
 85		if (ident.date_begin)
 86			*date = strtoul(ident.date_begin, NULL, 10);
 87	}
 88
 89	if (*line_end)
 90		return line_end + 1;
 91	else
 92		return line_end;
 93}
 94
 95#ifdef NO_ICONV
 96#define reencode(a, b, c)
 97#else
 98static const char *reencode(char **txt, const char *src_enc, const char *dst_enc)
 99{
100	char *tmp;
101
102	if (!txt)
103		return NULL;
104
105	if (!*txt || !src_enc || !dst_enc)
106		return *txt;
107
108	/* no encoding needed if src_enc equals dst_enc */
109	if (!strcasecmp(src_enc, dst_enc))
110		return *txt;
111
112	tmp = reencode_string(*txt, dst_enc, src_enc);
113	if (tmp) {
114		free(*txt);
115		*txt = tmp;
116	}
117	return *txt;
118}
119#endif
120
121static const char *next_header_line(const char *p)
122{
123	p = strchr(p, '\n');
124	if (!p)
125		return NULL;
126	return p + 1;
127}
128
129static int end_of_header(const char *p)
130{
131	return !p || (*p == '\n');
132}
133
134struct commitinfo *cgit_parse_commit(struct commit *commit)
135{
136	const int sha1hex_len = 40;
137	struct commitinfo *ret;
138	const char *p = get_cached_commit_buffer(commit, NULL);
139	const char *t;
140
141	ret = xcalloc(1, sizeof(struct commitinfo));
142	ret->commit = commit;
143
144	if (!p)
145		return ret;
146
147	if (!skip_prefix(p, "tree ", &p))
148		die("Bad commit: %s", sha1_to_hex(commit->object.sha1));
149	p += sha1hex_len + 1;
150
151	while (skip_prefix(p, "parent ", &p))
152		p += sha1hex_len + 1;
153
154	if (p && skip_prefix(p, "author ", &p)) {
155		p = parse_user(p, &ret->author, &ret->author_email,
156			&ret->author_date);
157	}
158
159	if (p && skip_prefix(p, "committer ", &p)) {
160		p = parse_user(p, &ret->committer, &ret->committer_email,
161			&ret->committer_date);
162	}
163
164	if (p && skip_prefix(p, "encoding ", &p)) {
165		t = strchr(p, '\n');
166		if (t) {
167			ret->msg_encoding = substr(p, t + 1);
168			p = t + 1;
169		}
170	}
171
172	if (!ret->msg_encoding)
173		ret->msg_encoding = xstrdup("UTF-8");
174
175	while (!end_of_header(p))
176		p = next_header_line(p);
177	while (p && *p == '\n')
178		p++;
179	if (!p)
180		return ret;
181
182	t = strchrnul(p, '\n');
183	ret->subject = substr(p, t);
184	while (*t == '\n')
185		t++;
186	ret->msg = xstrdup(t);
187
188	reencode(&ret->author, ret->msg_encoding, PAGE_ENCODING);
189	reencode(&ret->author_email, ret->msg_encoding, PAGE_ENCODING);
190	reencode(&ret->committer, ret->msg_encoding, PAGE_ENCODING);
191	reencode(&ret->committer_email, ret->msg_encoding, PAGE_ENCODING);
192	reencode(&ret->subject, ret->msg_encoding, PAGE_ENCODING);
193	reencode(&ret->msg, ret->msg_encoding, PAGE_ENCODING);
194
195	return ret;
196}
197
198struct taginfo *cgit_parse_tag(struct tag *tag)
199{
200	void *data;
201	enum object_type type;
202	unsigned long size;
203	const char *p;
204	struct taginfo *ret = NULL;
205
206	data = read_sha1_file(tag->object.sha1, &type, &size);
207	if (!data || type != OBJ_TAG)
208		goto cleanup;
209
210	ret = xcalloc(1, sizeof(struct taginfo));
211
212	for (p = data; !end_of_header(p); p = next_header_line(p)) {
213		if (skip_prefix(p, "tagger ", &p)) {
214			p = parse_user(p, &ret->tagger, &ret->tagger_email,
215				&ret->tagger_date);
216		}
217	}
218
219	while (p && *p == '\n')
220		p++;
221
222	if (p && *p)
223		ret->msg = xstrdup(p);
224
225cleanup:
226	free(data);
227	return ret;
228}