Context Navigation

source: nick.c @ 664bac3

Last change on this file since 664bac3 was fed4f76, checked in by dequis <dx@…>, at 2015-01-16T19:50:25Z

Fix UTF8 nick truncation issues

When nicks exceeded the length limit, they were cut at 24 bytes and that
sometimes left invalid utf8 at the end, which made the nick_ok()
validation fail and often broke those nicks completely.

This adds a truncate_utf8 function to cut the string at a safe place

Also, the method to deduplicate nicks when there's no more place to add
underscores was changed to add "_XX" at the end, where XX are two random
hex chars. The previous method in those cases was increasing the value
of the first character of the nick... which leads to silly and confusing
results (i.e. FacebookUser -> GacebookUser)

Property mode set to 100644

File size: 10.5 KB

Line
1	/********************************************************************\
2	* BitlBee -- An IRC to other IM-networks gateway *
3	* *
4	* Copyright 2002-2012 Wilmer van der Gaast and others *
5	\********************************************************************/
6
7	/* Some stuff to fetch, save and handle nicknames for your buddies */
8
9	/*
10	This program is free software; you can redistribute it and/or modify
11	it under the terms of the GNU General Public License as published by
12	the Free Software Foundation; either version 2 of the License, or
13	(at your option) any later version.
14
15	This program is distributed in the hope that it will be useful,
16	but WITHOUT ANY WARRANTY; without even the implied warranty of
17	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18	GNU General Public License for more details.
19
20	You should have received a copy of the GNU General Public License with
21	the Debian GNU/Linux distribution in /usr/share/common-licenses/GPL;
22	if not, write to the Free Software Foundation, Inc., 51 Franklin St.,
23	Fifth Floor, Boston, MA 02110-1301 USA
24	*/
25
26	#define BITLBEE_CORE
27	#include "bitlbee.h"
28
29	/* Character maps, _lc_[x] == _uc_[x] (but uppercase), according to the RFC's.
30	With one difference, we allow dashes. These are used to do uc/lc conversions
31	and strip invalid chars. */
32	static char *nick_lc_chars = "0123456789abcdefghijklmnopqrstuvwxyz{}^`-_\|";
33	static char *nick_uc_chars = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ[]~`-_\\";
34
35	/* Store handles in lower case and strip spaces, because AIM is braindead. */
36	static char clean_handle( const char orig )
37	{
38	char *new = g_malloc( strlen( orig ) + 1 );
39	int i = 0;
40
41	do {
42	if (*orig != ' ')
43	new[i++] = g_ascii_tolower( *orig );
44	}
45	while (*(orig++));
46
47	return new;
48	}
49
50	void nick_set_raw( account_t acc, const char handle, const char *nick )
51	{
52	char store_handle, store_nick = g_malloc( MAX_NICK_LENGTH + 1 );
53	irc_t irc = (irc_t ) acc->bee->ui_data;
54
55	store_handle = clean_handle( handle );
56	store_nick[MAX_NICK_LENGTH] = '\0';
57	strncpy( store_nick, nick, MAX_NICK_LENGTH );
58	nick_strip( irc, store_nick );
59
60	g_hash_table_replace( acc->nicks, store_handle, store_nick );
61	}
62
63	void nick_set( bee_user_t bu, const char nick )
64	{
65	nick_set_raw( bu->ic->acc, bu->handle, nick );
66	}
67
68	char nick_get( bee_user_t bu )
69	{
70	static char nick[MAX_NICK_LENGTH+1];
71	char store_handle, found_nick;
72	irc_t irc = (irc_t ) bu->bee->ui_data;
73
74	memset( nick, 0, MAX_NICK_LENGTH + 1 );
75
76	store_handle = clean_handle( bu->handle );
77	/* Find out if we stored a nick for this person already. If not, try
78	to generate a sane nick automatically. */
79	if( ( found_nick = g_hash_table_lookup( bu->ic->acc->nicks, store_handle ) ) )
80	{
81	strncpy( nick, found_nick, MAX_NICK_LENGTH );
82	}
83	else if( ( found_nick = nick_gen( bu ) ) )
84	{
85	strncpy( nick, found_nick, MAX_NICK_LENGTH );
86	g_free( found_nick );
87	}
88	else
89	{
90	/* Keep this fallback since nick_gen() can return NULL in some cases. */
91	char *s;
92
93	g_snprintf( nick, MAX_NICK_LENGTH, "%s", bu->handle );
94	if( ( s = strchr( nick, '@' ) ) )
95	while( *s )
96	*(s++) = 0;
97
98	nick_strip( irc, nick );
99	if( set_getbool( &bu->bee->set, "lcnicks" ) )
100	nick_lc( irc, nick );
101	}
102	g_free( store_handle );
103
104	/* Make sure the nick doesn't collide with an existing one by adding
105	underscores and that kind of stuff, if necessary. */
106	nick_dedupe( bu, nick );
107
108	return nick;
109	}
110
111	char nick_gen( bee_user_t bu )
112	{
113	gboolean ok = FALSE; /* Set to true once the nick contains something unique. */
114	GString *ret = g_string_sized_new( MAX_NICK_LENGTH + 1 );
115	char *rets;
116	irc_t irc = (irc_t ) bu->bee->ui_data;
117	char *fmt = set_getstr( &bu->ic->acc->set, "nick_format" ) ? :
118	set_getstr( &bu->bee->set, "nick_format" );
119
120	while( fmt && *fmt && ret->len < MAX_NICK_LENGTH )
121	{
122	char part = NULL, chop = '\0', asc = NULL, *s;
123	int len = INT_MAX;
124
125	if( *fmt != '%' )
126	{
127	g_string_append_c( ret, *fmt );
128	fmt ++;
129	continue;
130	}
131
132	fmt ++;
133	while( *fmt )
134	{
135	/* -char means chop off everything from char */
136	if( *fmt == '-' )
137	{
138	chop = fmt[1];
139	if( chop == '\0' )
140	{
141	g_string_free( ret, TRUE );
142	return NULL;
143	}
144	fmt += 2;
145	}
146	else if( g_ascii_isdigit( *fmt ) )
147	{
148	len = 0;
149	/* Grab a number. */
150	while( g_ascii_isdigit( *fmt ) )
151	len = len * 10 + ( *(fmt++) - '0' );
152	}
153	else if( g_strncasecmp( fmt, "nick", 4 ) == 0 )
154	{
155	part = bu->nick ? : bu->handle;
156	fmt += 4;
157	ok \|= TRUE;
158	break;
159	}
160	else if( g_strncasecmp( fmt, "handle", 6 ) == 0 )
161	{
162	part = bu->handle;
163	fmt += 6;
164	ok \|= TRUE;
165	break;
166	}
167	else if( g_strncasecmp( fmt, "full_name", 9 ) == 0 )
168	{
169	part = bu->fullname;
170	fmt += 9;
171	ok \|= part && *part;
172	break;
173	}
174	else if( g_strncasecmp( fmt, "first_name", 10 ) == 0 )
175	{
176	part = bu->fullname;
177	fmt += 10;
178	ok \|= part && *part;
179	chop = ' ';
180	break;
181	}
182	else if( g_strncasecmp( fmt, "group", 5 ) == 0 )
183	{
184	part = bu->group ? bu->group->name : NULL;
185	fmt += 5;
186	break;
187	}
188	else if( g_strncasecmp( fmt, "account", 7 ) == 0 )
189	{
190	part = bu->ic->acc->tag;
191	fmt += 7;
192	break;
193	}
194	else
195	{
196	g_string_free( ret, TRUE );
197	return NULL;
198	}
199	}
200
201	if( !part )
202	continue;
203
204	/* Credits to Josay_ in #bitlbee for this idea. //TRANSLIT
205	should do lossy/approximate conversions, so letters with
206	accents don't just get stripped. Note that it depends on
207	LC_CTYPE being set to something other than C/POSIX. */
208	if( !( irc && irc->status & IRC_UTF8_NICKS ) )
209	part = asc = g_convert_with_fallback( part, -1, "ASCII//TRANSLIT",
210	"UTF-8", "", NULL, NULL, NULL );
211
212	if( part && chop && ( s = strchr( part, chop ) ) )
213	len = MIN( len, s - part );
214
215	if( part )
216	{
217	if( len < INT_MAX )
218	g_string_append_len( ret, part, len );
219	else
220	g_string_append( ret, part );
221	}
222	g_free( asc );
223	}
224
225	rets = g_string_free( ret, FALSE );
226	if( ok && rets && *rets )
227	{
228	nick_strip( irc, rets );
229	truncate_utf8( rets, MAX_NICK_LENGTH );
230	return rets;
231	}
232	g_free( rets );
233	return NULL;
234	}
235
236	void nick_dedupe( bee_user_t *bu, char nick[MAX_NICK_LENGTH+1] )
237	{
238	irc_t irc = (irc_t) bu->bee->ui_data;
239	int inf_protection = 256;
240	irc_user_t *iu;
241
242	/* Now, find out if the nick is already in use at the moment, and make
243	subtle changes to make it unique. */
244	while( !nick_ok( irc, nick ) \|\|
245	( ( iu = irc_user_by_name( irc, nick ) ) && iu->bu != bu ) )
246	{
247	if( strlen( nick ) < ( MAX_NICK_LENGTH - 1 ) )
248	{
249	nick[strlen(nick)+1] = 0;
250	nick[strlen(nick)] = '_';
251	}
252	else
253	{
254	/* We've got no more space for underscores,
255	so truncate it and replace the last three
256	chars with a random "_XX" suffix */
257	int len = truncate_utf8( nick, MAX_NICK_LENGTH - 3 );
258	nick[len] = '_';
259	g_snprintf(nick + len + 1, 3, "%2x", rand() );
260	}
261
262	if( inf_protection-- == 0 )
263	{
264	g_snprintf( nick, MAX_NICK_LENGTH + 1, "xx%x", rand() );
265
266	irc_rootmsg( irc, "Warning: Something went wrong while trying "
267	"to generate a nickname for contact %s on %s.",
268	bu->handle, bu->ic->acc->tag );
269	irc_rootmsg( irc, "This might be a bug in BitlBee, or the result "
270	"of a faulty nick_format setting. Will use %s "
271	"instead.", nick );
272
273	break;
274	}
275	}
276	}
277
278	/* Just check if there is a nickname set for this buddy or if we'd have to
279	generate one. */
280	int nick_saved( bee_user_t *bu )
281	{
282	char store_handle, found;
283
284	store_handle = clean_handle( bu->handle );
285	found = g_hash_table_lookup( bu->ic->acc->nicks, store_handle );
286	g_free( store_handle );
287
288	return found != NULL;
289	}
290
291	void nick_del( bee_user_t *bu )
292	{
293	g_hash_table_remove( bu->ic->acc->nicks, bu->handle );
294	}
295
296
297	void nick_strip( irc_t irc, char nick )
298	{
299	int len = 0;
300
301	if( irc && ( irc->status & IRC_UTF8_NICKS ) )
302	{
303	gunichar c;
304	char p = nick, n, tmp[strlen(nick)+1];
305
306	while( p && *p )
307	{
308	c = g_utf8_get_char_validated( p, -1 );
309	n = g_utf8_find_next_char( p, NULL );
310
311	if( ( c < 0x7f && !( strchr( nick_lc_chars, c ) \|\|
312	strchr( nick_uc_chars, c ) ) ) \|\|
313	!g_unichar_isgraph( c ) )
314	{
315	strcpy( tmp, n );
316	strcpy( p, tmp );
317	}
318	else
319	p = n;
320	}
321	if( p )
322	len = p - nick;
323	}
324	else
325	{
326	int i;
327
328	for( i = len = 0; nick[i] && len < MAX_NICK_LENGTH; i++ )
329	{
330	if( strchr( nick_lc_chars, nick[i] ) \|\|
331	strchr( nick_uc_chars, nick[i] ) )
332	{
333	nick[len] = nick[i];
334	len++;
335	}
336	}
337	}
338	if( g_ascii_isdigit( nick[0] ) )
339	{
340	char *orig;
341
342	/* First character of a nick can't be a digit, so insert an
343	underscore if necessary. */
344	orig = g_strdup( nick );
345	g_snprintf( nick, MAX_NICK_LENGTH, "_%s", orig );
346	g_free( orig );
347	len ++;
348	}
349	while( len <= MAX_NICK_LENGTH )
350	nick[len++] = '\0';
351	}
352
353	gboolean nick_ok( irc_t irc, const char nick )
354	{
355	const char *s;
356
357	/* Empty/long nicks are not allowed, nor numbers at [0] */
358	if( !*nick \|\| g_ascii_isdigit( nick[0] ) \|\| strlen( nick ) > MAX_NICK_LENGTH )
359	return 0;
360
361	if( irc && ( irc->status & IRC_UTF8_NICKS ) )
362	{
363	gunichar c;
364	const char p = nick, n;
365
366	while( p && *p )
367	{
368	c = g_utf8_get_char_validated( p, -1 );
369	n = g_utf8_find_next_char( p, NULL );
370
371	if( ( c < 0x7f && !( strchr( nick_lc_chars, c ) \|\|
372	strchr( nick_uc_chars, c ) ) ) \|\|
373	!g_unichar_isgraph( c ) )
374	{
375	return FALSE;
376	}
377	p = n;
378	}
379	}
380	else
381	{
382	for( s = nick; *s; s ++ )
383	if( !strchr( nick_lc_chars, s ) && !strchr( nick_uc_chars, s ) )
384	return FALSE;
385	}
386
387	return TRUE;
388	}
389
390	int nick_lc( irc_t irc, char nick )
391	{
392	static char tab[128] = { 0 };
393	int i;
394
395	if( tab['A'] == 0 )
396	for( i = 0; nick_lc_chars[i]; i ++ )
397	{
398	tab[(int)nick_uc_chars[i]] = nick_lc_chars[i];
399	tab[(int)nick_lc_chars[i]] = nick_lc_chars[i];
400	}
401
402	if( irc && ( irc->status & IRC_UTF8_NICKS ) )
403	{
404	gchar *down = g_utf8_strdown( nick, -1 );
405	if( strlen( down ) > strlen( nick ) )
406	{
407	truncate_utf8( down, strlen(nick) );
408	}
409	strcpy( nick, down );
410	g_free( down );
411	}
412
413	for( i = 0; nick[i]; i ++ )
414	if( ((guchar)nick[i]) < 0x7f )
415	nick[i] = tab[(guchar)nick[i]];
416
417	return nick_ok( irc, nick );
418	}
419
420	int nick_cmp( irc_t irc, const char a, const char *b )
421	{
422	char aa[1024] = "", bb[1024] = "";
423
424	strncpy( aa, a, sizeof( aa ) - 1 );
425	strncpy( bb, b, sizeof( bb ) - 1 );
426	if( nick_lc( irc, aa ) && nick_lc( irc, bb ) )
427	{
428	return( strcmp( aa, bb ) );
429	}
430	else
431	{
432	return( -1 ); /* Hmm... Not a clear answer.. :-/ */
433	}
434	}

Note: See TracBrowser for help on using the repository browser.

Context Navigation

source: nick.c @ 664bac3

Download in other formats: