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