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