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 : #include <assert.h>
26 :
27 : #include "gpg.h"
28 : #include "status.h"
29 : #include "iobuf.h"
30 : #include "util.h"
31 : #include "filter.h"
32 : #include "i18n.h"
33 : #include "options.h"
34 : #include "status.h"
35 :
36 : #ifdef HAVE_DOSISH_SYSTEM
37 : #define LF "\r\n"
38 : #else
39 : #define LF "\n"
40 : #endif
41 :
42 : #define MAX_LINELEN 19995 /* a little bit smaller than in armor.c */
43 : /* to make sure that a warning is displayed while */
44 : /* creating a message */
45 :
46 : static unsigned
47 5228 : len_without_trailing_chars( byte *line, unsigned len, const char *trimchars )
48 : {
49 : byte *p, *mark;
50 : unsigned n;
51 :
52 208143 : for(mark=NULL, p=line, n=0; n < len; n++, p++ ) {
53 202915 : if( strchr( trimchars, *p ) ) {
54 38632 : if( !mark )
55 32726 : mark = p;
56 : }
57 : else
58 164283 : mark = NULL;
59 : }
60 :
61 5228 : return mark? (mark - line) : len;
62 : }
63 :
64 :
65 : static int
66 6 : standard( text_filter_context_t *tfx, IOBUF a,
67 : byte *buf, size_t size, size_t *ret_len)
68 : {
69 6 : int rc=0;
70 6 : size_t len = 0;
71 : unsigned maxlen;
72 :
73 6 : assert( size > 10 );
74 6 : size -= 2; /* reserve 2 bytes to append CR,LF */
75 85 : while( !rc && len < size ) {
76 : int lf_seen;
77 :
78 2940 : while( len < size && tfx->buffer_pos < tfx->buffer_len )
79 2782 : buf[len++] = tfx->buffer[tfx->buffer_pos++];
80 79 : if( len >= size )
81 0 : continue;
82 :
83 : /* read the next line */
84 79 : maxlen = MAX_LINELEN;
85 79 : tfx->buffer_pos = 0;
86 79 : tfx->buffer_len = iobuf_read_line( a, &tfx->buffer,
87 : &tfx->buffer_size, &maxlen );
88 79 : if( !maxlen )
89 0 : tfx->truncated++;
90 79 : if( !tfx->buffer_len ) {
91 6 : if( !len )
92 3 : rc = -1; /* eof */
93 6 : break;
94 : }
95 73 : lf_seen = tfx->buffer[tfx->buffer_len-1] == '\n';
96 :
97 : /* The story behind this is that 2440 says that textmode
98 : hashes should canonicalize line endings to CRLF and remove
99 : spaces and tabs. 2440bis-12 says to just canonicalize to
100 : CRLF. 1.4.0 was released using the bis-12 behavior, but it
101 : was discovered that many mail clients do not canonicalize
102 : PGP/MIME signature text appropriately (and were relying on
103 : GnuPG to handle trailing spaces). So, we default to the
104 : 2440 behavior, but use the 2440bis-12 behavior if the user
105 : specifies --no-rfc2440-text. The default will be changed
106 : at some point in the future when the mail clients have been
107 : upgraded. Aside from PGP/MIME and broken mail clients,
108 : this makes no difference to any signatures in the real
109 : world except for a textmode detached signature. PGP always
110 : used the 2440bis-12 behavior (ignoring 2440 itself), so
111 : this actually makes us compatible with PGP textmode
112 : detached signatures for the first time. */
113 73 : if(opt.rfc2440_text)
114 0 : tfx->buffer_len=trim_trailing_chars(tfx->buffer,tfx->buffer_len,
115 : " \t\r\n");
116 : else
117 73 : tfx->buffer_len=trim_trailing_chars(tfx->buffer,tfx->buffer_len,
118 : "\r\n");
119 :
120 73 : if( lf_seen ) {
121 72 : tfx->buffer[tfx->buffer_len++] = '\r';
122 72 : tfx->buffer[tfx->buffer_len++] = '\n';
123 : }
124 : }
125 6 : *ret_len = len;
126 6 : return rc;
127 : }
128 :
129 :
130 : /****************
131 : * The filter is used to make canonical text: Lines are terminated by
132 : * CR, LF, trailing white spaces are removed.
133 : */
134 : int
135 12 : text_filter( void *opaque, int control,
136 : IOBUF a, byte *buf, size_t *ret_len)
137 : {
138 12 : size_t size = *ret_len;
139 12 : text_filter_context_t *tfx = opaque;
140 12 : int rc=0;
141 :
142 12 : if( control == IOBUFCTRL_UNDERFLOW ) {
143 6 : rc = standard( tfx, a, buf, size, ret_len );
144 : }
145 6 : else if( control == IOBUFCTRL_FREE ) {
146 3 : if( tfx->truncated )
147 0 : log_error(_("can't handle text lines longer than %d characters\n"),
148 : MAX_LINELEN );
149 3 : xfree( tfx->buffer );
150 3 : tfx->buffer = NULL;
151 : }
152 3 : else if( control == IOBUFCTRL_DESC )
153 0 : *(char**)buf = "text_filter";
154 12 : return rc;
155 : }
156 :
157 :
158 : /****************
159 : * Copy data from INP to OUT and do some escaping if requested.
160 : * md is updated as required by rfc2440
161 : */
162 : int
163 10 : copy_clearsig_text( IOBUF out, IOBUF inp, gcry_md_hd_t md,
164 : int escape_dash, int escape_from)
165 : {
166 : unsigned int maxlen;
167 10 : byte *buffer = NULL; /* malloced buffer */
168 : unsigned int bufsize; /* and size of this buffer */
169 : unsigned int n;
170 10 : int truncated = 0;
171 10 : int pending_lf = 0;
172 :
173 10 : if( !escape_dash )
174 1 : escape_from = 0;
175 :
176 10 : write_status_begin_signing (md);
177 :
178 : for(;;) {
179 5249 : maxlen = MAX_LINELEN;
180 5249 : n = iobuf_read_line( inp, &buffer, &bufsize, &maxlen );
181 5249 : if( !maxlen )
182 0 : truncated++;
183 :
184 5249 : if( !n )
185 10 : break; /* read_line has returned eof */
186 :
187 : /* update the message digest */
188 5239 : if( escape_dash ) {
189 5228 : if( pending_lf ) {
190 5220 : gcry_md_putc ( md, '\r' );
191 5220 : gcry_md_putc ( md, '\n' );
192 : }
193 5228 : gcry_md_write ( md, buffer,
194 5228 : len_without_trailing_chars (buffer, n, " \t\r\n"));
195 : }
196 : else
197 11 : gcry_md_write ( md, buffer, n );
198 5239 : pending_lf = buffer[n-1] == '\n';
199 :
200 : /* write the output */
201 5239 : if( ( escape_dash && *buffer == '-')
202 5231 : || ( escape_from && n > 4 && !memcmp(buffer, "From ", 5 ) ) ) {
203 8 : iobuf_put( out, '-' );
204 8 : iobuf_put( out, ' ' );
205 : }
206 :
207 : #if 0 /*defined(HAVE_DOSISH_SYSTEM)*/
208 : /* We don't use this anymore because my interpretation of rfc2440 7.1
209 : * is that there is no conversion needed. If one decides to
210 : * clearsign a unix file on a DOS box he will get a mixed line endings.
211 : * If at some point it turns out, that a conversion is a nice feature
212 : * we can make an option out of it.
213 : */
214 : /* make sure the lines do end in CR,LF */
215 : if( n > 1 && ( (buffer[n-2] == '\r' && buffer[n-1] == '\n' )
216 : || (buffer[n-2] == '\n' && buffer[n-1] == '\r'))) {
217 : iobuf_write( out, buffer, n-2 );
218 : iobuf_put( out, '\r');
219 : iobuf_put( out, '\n');
220 : }
221 : else if( n && buffer[n-1] == '\n' ) {
222 : iobuf_write( out, buffer, n-1 );
223 : iobuf_put( out, '\r');
224 : iobuf_put( out, '\n');
225 : }
226 : else
227 : iobuf_write( out, buffer, n );
228 :
229 : #else
230 5239 : iobuf_write( out, buffer, n );
231 : #endif
232 5239 : }
233 :
234 : /* at eof */
235 10 : if( !pending_lf ) { /* make sure that the file ends with a LF */
236 3 : iobuf_writestr( out, LF );
237 3 : if( !escape_dash )
238 0 : gcry_md_putc( md, '\n' );
239 : }
240 :
241 10 : if( truncated )
242 0 : log_info(_("input line longer than %d characters\n"), MAX_LINELEN );
243 :
244 10 : return 0; /* okay */
245 : }
|
