Line data Source code
1 : /* textfilter.c
2 : * Copyright (C) 1998, 1999, 2000, 2001, 2004 Free Software Foundation, Inc.
3 : *
4 : * This file is part of GnuPG.
5 : *
6 : * GnuPG is free software; you can redistribute it and/or modify
7 : * it under the terms of the GNU General Public License as published by
8 : * the Free Software Foundation; either version 3 of the License, or
9 : * (at your option) any later version.
10 : *
11 : * GnuPG is distributed in the hope that it will be useful,
12 : * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 : * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 : * GNU General Public License for more details.
15 : *
16 : * You should have received a copy of the GNU General Public License
17 : * along with this program; if not, see <http://www.gnu.org/licenses/>.
18 : */
19 :
20 : #include <config.h>
21 : #include <stdio.h>
22 : #include <stdlib.h>
23 : #include <string.h>
24 : #include <errno.h>
25 :
26 : #include "gpg.h"
27 : #include "status.h"
28 : #include "iobuf.h"
29 : #include "util.h"
30 : #include "filter.h"
31 : #include "i18n.h"
32 : #include "options.h"
33 : #include "status.h"
34 :
35 : #ifdef HAVE_DOSISH_SYSTEM
36 : #define LF "\r\n"
37 : #else
38 : #define LF "\n"
39 : #endif
40 :
41 : #define MAX_LINELEN 19995 /* a little bit smaller than in armor.c */
42 : /* to make sure that a warning is displayed while */
43 : /* creating a message */
44 :
45 : static unsigned
46 5267 : len_without_trailing_chars( byte *line, unsigned len, const char *trimchars )
47 : {
48 : byte *p, *mark;
49 : unsigned n;
50 :
51 209895 : for(mark=NULL, p=line, n=0; n < len; n++, p++ ) {
52 204628 : if( strchr( trimchars, *p ) ) {
53 38965 : if( !mark )
54 33014 : mark = p;
55 : }
56 : else
57 165663 : mark = NULL;
58 : }
59 :
60 5267 : return mark? (mark - line) : len;
61 : }
62 :
63 :
64 : static int
65 6 : standard( text_filter_context_t *tfx, IOBUF a,
66 : byte *buf, size_t size, size_t *ret_len)
67 : {
68 6 : int rc=0;
69 6 : size_t len = 0;
70 : unsigned maxlen;
71 :
72 6 : log_assert( size > 10 );
73 6 : size -= 2; /* reserve 2 bytes to append CR,LF */
74 85 : while( !rc && len < size ) {
75 : int lf_seen;
76 :
77 2940 : while( len < size && tfx->buffer_pos < tfx->buffer_len )
78 2782 : buf[len++] = tfx->buffer[tfx->buffer_pos++];
79 79 : if( len >= size )
80 0 : continue;
81 :
82 : /* read the next line */
83 79 : maxlen = MAX_LINELEN;
84 79 : tfx->buffer_pos = 0;
85 79 : tfx->buffer_len = iobuf_read_line( a, &tfx->buffer,
86 : &tfx->buffer_size, &maxlen );
87 79 : if( !maxlen )
88 0 : tfx->truncated++;
89 79 : if( !tfx->buffer_len ) {
90 6 : if( !len )
91 3 : rc = -1; /* eof */
92 6 : break;
93 : }
94 73 : lf_seen = tfx->buffer[tfx->buffer_len-1] == '\n';
95 :
96 : /* The story behind this is that 2440 says that textmode
97 : hashes should canonicalize line endings to CRLF and remove
98 : spaces and tabs. 2440bis-12 says to just canonicalize to
99 : CRLF. 1.4.0 was released using the bis-12 behavior, but it
100 : was discovered that many mail clients do not canonicalize
101 : PGP/MIME signature text appropriately (and were relying on
102 : GnuPG to handle trailing spaces). So, we default to the
103 : 2440 behavior, but use the 2440bis-12 behavior if the user
104 : specifies --no-rfc2440-text. The default will be changed
105 : at some point in the future when the mail clients have been
106 : upgraded. Aside from PGP/MIME and broken mail clients,
107 : this makes no difference to any signatures in the real
108 : world except for a textmode detached signature. PGP always
109 : used the 2440bis-12 behavior (ignoring 2440 itself), so
110 : this actually makes us compatible with PGP textmode
111 : detached signatures for the first time. */
112 73 : if(opt.rfc2440_text)
113 0 : tfx->buffer_len=trim_trailing_chars(tfx->buffer,tfx->buffer_len,
114 : " \t\r\n");
115 : else
116 73 : tfx->buffer_len=trim_trailing_chars(tfx->buffer,tfx->buffer_len,
117 : "\r\n");
118 :
119 73 : if( lf_seen ) {
120 72 : tfx->buffer[tfx->buffer_len++] = '\r';
121 72 : tfx->buffer[tfx->buffer_len++] = '\n';
122 : }
123 : }
124 6 : *ret_len = len;
125 6 : return rc;
126 : }
127 :
128 :
129 : /****************
130 : * The filter is used to make canonical text: Lines are terminated by
131 : * CR, LF, trailing white spaces are removed.
132 : */
133 : int
134 12 : text_filter( void *opaque, int control,
135 : IOBUF a, byte *buf, size_t *ret_len)
136 : {
137 12 : size_t size = *ret_len;
138 12 : text_filter_context_t *tfx = opaque;
139 12 : int rc=0;
140 :
141 12 : if( control == IOBUFCTRL_UNDERFLOW ) {
142 6 : rc = standard( tfx, a, buf, size, ret_len );
143 : }
144 6 : else if( control == IOBUFCTRL_FREE ) {
145 3 : if( tfx->truncated )
146 0 : log_error(_("can't handle text lines longer than %d characters\n"),
147 : MAX_LINELEN );
148 3 : xfree( tfx->buffer );
149 3 : tfx->buffer = NULL;
150 : }
151 3 : else if( control == IOBUFCTRL_DESC )
152 0 : mem2str (buf, "text_filter", *ret_len);
153 12 : return rc;
154 : }
155 :
156 :
157 : /****************
158 : * Copy data from INP to OUT and do some escaping if requested.
159 : * md is updated as required by rfc2440
160 : */
161 : int
162 10 : copy_clearsig_text( IOBUF out, IOBUF inp, gcry_md_hd_t md,
163 : int escape_dash, int escape_from)
164 : {
165 : unsigned int maxlen;
166 10 : byte *buffer = NULL; /* malloced buffer */
167 : unsigned int bufsize; /* and size of this buffer */
168 : unsigned int n;
169 10 : int truncated = 0;
170 10 : int pending_lf = 0;
171 :
172 10 : if( !escape_dash )
173 1 : escape_from = 0;
174 :
175 10 : write_status_begin_signing (md);
176 :
177 : for(;;) {
178 5288 : maxlen = MAX_LINELEN;
179 5288 : n = iobuf_read_line( inp, &buffer, &bufsize, &maxlen );
180 5288 : if( !maxlen )
181 0 : truncated++;
182 :
183 5288 : if( !n )
184 10 : break; /* read_line has returned eof */
185 :
186 : /* update the message digest */
187 5278 : if( escape_dash ) {
188 5267 : if( pending_lf ) {
189 5259 : gcry_md_putc ( md, '\r' );
190 5259 : gcry_md_putc ( md, '\n' );
191 : }
192 5267 : gcry_md_write ( md, buffer,
193 5267 : len_without_trailing_chars (buffer, n, " \t\r\n"));
194 : }
195 : else
196 11 : gcry_md_write ( md, buffer, n );
197 5278 : pending_lf = buffer[n-1] == '\n';
198 :
199 : /* write the output */
200 5278 : if( ( escape_dash && *buffer == '-')
201 5270 : || ( escape_from && n > 4 && !memcmp(buffer, "From ", 5 ) ) ) {
202 8 : iobuf_put( out, '-' );
203 8 : iobuf_put( out, ' ' );
204 : }
205 :
206 : #if 0 /*defined(HAVE_DOSISH_SYSTEM)*/
207 : /* We don't use this anymore because my interpretation of rfc2440 7.1
208 : * is that there is no conversion needed. If one decides to
209 : * clearsign a unix file on a DOS box he will get a mixed line endings.
210 : * If at some point it turns out, that a conversion is a nice feature
211 : * we can make an option out of it.
212 : */
213 : /* make sure the lines do end in CR,LF */
214 : if( n > 1 && ( (buffer[n-2] == '\r' && buffer[n-1] == '\n' )
215 : || (buffer[n-2] == '\n' && buffer[n-1] == '\r'))) {
216 : iobuf_write( out, buffer, n-2 );
217 : iobuf_put( out, '\r');
218 : iobuf_put( out, '\n');
219 : }
220 : else if( n && buffer[n-1] == '\n' ) {
221 : iobuf_write( out, buffer, n-1 );
222 : iobuf_put( out, '\r');
223 : iobuf_put( out, '\n');
224 : }
225 : else
226 : iobuf_write( out, buffer, n );
227 :
228 : #else
229 5278 : iobuf_write( out, buffer, n );
230 : #endif
231 5278 : }
232 :
233 : /* at eof */
234 10 : if( !pending_lf ) { /* make sure that the file ends with a LF */
235 3 : iobuf_writestr( out, LF );
236 3 : if( !escape_dash )
237 0 : gcry_md_putc( md, '\n' );
238 : }
239 :
240 10 : if( truncated )
241 0 : log_info(_("input line longer than %d characters\n"), MAX_LINELEN );
242 :
243 10 : xfree (buffer);
244 10 : return 0; /* okay */
245 : }
|