Eric Andersen | 9670083 | 2000-09-04 15:15:55 +0000 | [diff] [blame] | 1 | /* vi: set sw=4 ts=4: */ |
| 2 | /* |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 3 | * wget - retrieve a file using HTTP or FTP |
Eric Andersen | 9670083 | 2000-09-04 15:15:55 +0000 | [diff] [blame] | 4 | * |
Eric Andersen | 4e573f4 | 2000-11-14 23:29:24 +0000 | [diff] [blame] | 5 | * Chip Rosenthal Covad Communications <chip@laserlink.net> |
Denys Vlasenko | 0ef64bd | 2010-08-16 20:14:46 +0200 | [diff] [blame] | 6 | * Licensed under GPLv2, see file LICENSE in this source tree. |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 7 | * |
| 8 | * Copyright (C) 2010 Bradley M. Kuhn <bkuhn@ebb.org> |
Denys Vlasenko | fb132e4 | 2010-10-29 11:46:52 +0200 | [diff] [blame] | 9 | * Kuhn's copyrights are licensed GPLv2-or-later. File as a whole remains GPLv2. |
Eric Andersen | 9670083 | 2000-09-04 15:15:55 +0000 | [diff] [blame] | 10 | */ |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 11 | //config:config WGET |
Denys Vlasenko | 4eed2c6 | 2017-07-18 22:01:24 +0200 | [diff] [blame] | 12 | //config: bool "wget (35 kb)" |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 13 | //config: default y |
| 14 | //config: help |
Denys Vlasenko | 72089cf | 2017-07-21 09:50:55 +0200 | [diff] [blame] | 15 | //config: wget is a utility for non-interactive download of files from HTTP |
| 16 | //config: and FTP servers. |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 17 | //config: |
Denys Vlasenko | f560422 | 2017-01-10 14:58:54 +0100 | [diff] [blame] | 18 | //config:config FEATURE_WGET_LONG_OPTIONS |
| 19 | //config: bool "Enable long options" |
| 20 | //config: default y |
| 21 | //config: depends on WGET && LONG_OPTS |
| 22 | //config: |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 23 | //config:config FEATURE_WGET_STATUSBAR |
Denys Vlasenko | f560422 | 2017-01-10 14:58:54 +0100 | [diff] [blame] | 24 | //config: bool "Enable progress bar (+2k)" |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 25 | //config: default y |
| 26 | //config: depends on WGET |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 27 | //config: |
| 28 | //config:config FEATURE_WGET_AUTHENTICATION |
| 29 | //config: bool "Enable HTTP authentication" |
| 30 | //config: default y |
| 31 | //config: depends on WGET |
| 32 | //config: help |
Denys Vlasenko | 72089cf | 2017-07-21 09:50:55 +0200 | [diff] [blame] | 33 | //config: Support authenticated HTTP transfers. |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 34 | //config: |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 35 | //config:config FEATURE_WGET_TIMEOUT |
| 36 | //config: bool "Enable timeout option -T SEC" |
| 37 | //config: default y |
| 38 | //config: depends on WGET |
| 39 | //config: help |
Denys Vlasenko | 72089cf | 2017-07-21 09:50:55 +0200 | [diff] [blame] | 40 | //config: Supports network read and connect timeouts for wget, |
| 41 | //config: so that wget will give up and timeout, through the -T |
| 42 | //config: command line option. |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 43 | //config: |
Denys Vlasenko | 72089cf | 2017-07-21 09:50:55 +0200 | [diff] [blame] | 44 | //config: Currently only connect and network data read timeout are |
| 45 | //config: supported (i.e., timeout is not applied to the DNS query). When |
| 46 | //config: FEATURE_WGET_LONG_OPTIONS is also enabled, the --timeout option |
| 47 | //config: will work in addition to -T. |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 48 | //config: |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 49 | //config:config FEATURE_WGET_HTTPS |
| 50 | //config: bool "Support HTTPS using internal TLS code" |
Denys Vlasenko | 403f299 | 2018-02-06 15:15:08 +0100 | [diff] [blame] | 51 | //it also enables FTPS support, but it's not well tested yet |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 52 | //config: default y |
| 53 | //config: depends on WGET |
| 54 | //config: select TLS |
| 55 | //config: help |
Denys Vlasenko | 72089cf | 2017-07-21 09:50:55 +0200 | [diff] [blame] | 56 | //config: wget will use internal TLS code to connect to https:// URLs. |
| 57 | //config: Note: |
| 58 | //config: On NOMMU machines, ssl_helper applet should be available |
| 59 | //config: in the $PATH for this to work. Make sure to select that applet. |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 60 | //config: |
Denys Vlasenko | 72089cf | 2017-07-21 09:50:55 +0200 | [diff] [blame] | 61 | //config: Note: currently, TLS code only makes TLS I/O work, it |
| 62 | //config: does *not* check that the peer is who it claims to be, etc. |
| 63 | //config: IOW: it uses peer-supplied public keys to establish encryption |
| 64 | //config: and signing keys, then encrypts and signs outgoing data and |
| 65 | //config: decrypts incoming data. |
| 66 | //config: It does not check signature hashes on the incoming data: |
| 67 | //config: this means that attackers manipulating TCP packets can |
| 68 | //config: send altered data and we unknowingly receive garbage. |
| 69 | //config: (This check might be relatively easy to add). |
| 70 | //config: It does not check public key's certificate: |
| 71 | //config: this means that the peer may be an attacker impersonating |
| 72 | //config: the server we think we are talking to. |
Denys Vlasenko | 67f6db6 | 2017-01-30 16:27:37 +0100 | [diff] [blame] | 73 | //config: |
Denys Vlasenko | 72089cf | 2017-07-21 09:50:55 +0200 | [diff] [blame] | 74 | //config: If you think this is unacceptable, consider this. As more and more |
| 75 | //config: servers switch to HTTPS-only operation, without such "crippled" |
| 76 | //config: TLS code it is *impossible* to simply download a kernel source |
| 77 | //config: from kernel.org. Which can in real world translate into |
| 78 | //config: "my small automatic tooling to build cross-compilers from sources |
| 79 | //config: no longer works, I need to additionally keep a local copy |
| 80 | //config: of ~4 megabyte source tarball of a SSL library and ~2 megabyte |
| 81 | //config: source of wget, need to compile and built both before I can |
| 82 | //config: download anything. All this despite the fact that the build |
| 83 | //config: is done in a QEMU sandbox on a machine with absolutely nothing |
| 84 | //config: worth stealing, so I don't care if someone would go to a lot |
| 85 | //config: of trouble to intercept my HTTPS download to send me an altered |
| 86 | //config: kernel tarball". |
Denys Vlasenko | 67f6db6 | 2017-01-30 16:27:37 +0100 | [diff] [blame] | 87 | //config: |
Denys Vlasenko | 72089cf | 2017-07-21 09:50:55 +0200 | [diff] [blame] | 88 | //config: If you still think this is unacceptable, send patches. |
Denys Vlasenko | 67f6db6 | 2017-01-30 16:27:37 +0100 | [diff] [blame] | 89 | //config: |
Denys Vlasenko | 72089cf | 2017-07-21 09:50:55 +0200 | [diff] [blame] | 90 | //config: If you still think this is unacceptable, do not want to send |
| 91 | //config: patches, but do want to waste bandwidth expaining how wrong |
| 92 | //config: it is, you will be ignored. |
Denys Vlasenko | 67f6db6 | 2017-01-30 16:27:37 +0100 | [diff] [blame] | 93 | //config: |
Denys Vlasenko | 2007ef5 | 2015-10-07 02:40:53 +0200 | [diff] [blame] | 94 | //config:config FEATURE_WGET_OPENSSL |
| 95 | //config: bool "Try to connect to HTTPS using openssl" |
| 96 | //config: default y |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 97 | //config: depends on WGET |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 98 | //config: help |
Denys Vlasenko | 72089cf | 2017-07-21 09:50:55 +0200 | [diff] [blame] | 99 | //config: Try to use openssl to handle HTTPS. |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 100 | //config: |
Denys Vlasenko | 72089cf | 2017-07-21 09:50:55 +0200 | [diff] [blame] | 101 | //config: OpenSSL has a simple SSL client for debug purposes. |
| 102 | //config: If you select this option, wget will effectively run: |
| 103 | //config: "openssl s_client -quiet -connect hostname:443 |
| 104 | //config: -servername hostname 2>/dev/null" and pipe its data |
| 105 | //config: through it. -servername is not used if hostname is numeric. |
| 106 | //config: Note inconvenient API: host resolution is done twice, |
| 107 | //config: and there is no guarantee openssl's idea of IPv6 address |
| 108 | //config: format is the same as ours. |
| 109 | //config: Another problem is that s_client prints debug information |
| 110 | //config: to stderr, and it needs to be suppressed. This means |
| 111 | //config: all error messages get suppressed too. |
| 112 | //config: openssl is also a big binary, often dynamically linked |
| 113 | //config: against ~15 libraries. |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 114 | //config: |
Denys Vlasenko | 72089cf | 2017-07-21 09:50:55 +0200 | [diff] [blame] | 115 | //config: If openssl can't be executed, internal TLS code will be used |
| 116 | //config: (if you enabled it); if openssl can be executed but fails later, |
| 117 | //config: wget can't detect this, and download will fail. |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 118 | |
| 119 | //applet:IF_WGET(APPLET(wget, BB_DIR_USR_BIN, BB_SUID_DROP)) |
| 120 | |
| 121 | //kbuild:lib-$(CONFIG_WGET) += wget.o |
| 122 | |
Denys Vlasenko | e2e55b0 | 2011-03-21 00:37:05 +0100 | [diff] [blame] | 123 | //usage:#define wget_trivial_usage |
| 124 | //usage: IF_FEATURE_WGET_LONG_OPTIONS( |
Denys Vlasenko | 2972e2c | 2016-10-04 04:23:09 +0200 | [diff] [blame] | 125 | //usage: "[-c|--continue] [--spider] [-q|--quiet] [-O|--output-document FILE]\n" |
Denys Vlasenko | e2e55b0 | 2011-03-21 00:37:05 +0100 | [diff] [blame] | 126 | //usage: " [--header 'header: value'] [-Y|--proxy on/off] [-P DIR]\n" |
Vladimir Dronnikov | f5abc78 | 2012-06-13 17:29:41 +0200 | [diff] [blame] | 127 | /* Since we ignore these opts, we don't show them in --help */ |
Denys Vlasenko | 92e1b08 | 2015-10-20 21:51:52 +0200 | [diff] [blame] | 128 | /* //usage: " [--no-check-certificate] [--no-cache] [--passive-ftp] [-t TRIES]" */ |
| 129 | /* //usage: " [-nv] [-nc] [-nH] [-np]" */ |
Denys Vlasenko | dff9fef | 2017-01-24 21:41:43 +0100 | [diff] [blame] | 130 | //usage: " [-S|--server-response] [-U|--user-agent AGENT]" IF_FEATURE_WGET_TIMEOUT(" [-T SEC]") " URL..." |
Denys Vlasenko | e2e55b0 | 2011-03-21 00:37:05 +0100 | [diff] [blame] | 131 | //usage: ) |
| 132 | //usage: IF_NOT_FEATURE_WGET_LONG_OPTIONS( |
Denys Vlasenko | dff9fef | 2017-01-24 21:41:43 +0100 | [diff] [blame] | 133 | //usage: "[-cq] [-O FILE] [-Y on/off] [-P DIR] [-S] [-U AGENT]" |
Denys Vlasenko | e2e55b0 | 2011-03-21 00:37:05 +0100 | [diff] [blame] | 134 | //usage: IF_FEATURE_WGET_TIMEOUT(" [-T SEC]") " URL..." |
| 135 | //usage: ) |
| 136 | //usage:#define wget_full_usage "\n\n" |
| 137 | //usage: "Retrieve files via HTTP or FTP\n" |
Denys Vlasenko | 2972e2c | 2016-10-04 04:23:09 +0200 | [diff] [blame] | 138 | //usage: IF_FEATURE_WGET_LONG_OPTIONS( |
Denys Vlasenko | dff9fef | 2017-01-24 21:41:43 +0100 | [diff] [blame] | 139 | //usage: "\n --spider Only check URL existence: $? is 0 if exists" |
Denys Vlasenko | e2e55b0 | 2011-03-21 00:37:05 +0100 | [diff] [blame] | 140 | //usage: ) |
Denys Vlasenko | 2972e2c | 2016-10-04 04:23:09 +0200 | [diff] [blame] | 141 | //usage: "\n -c Continue retrieval of aborted transfer" |
| 142 | //usage: "\n -q Quiet" |
| 143 | //usage: "\n -P DIR Save to DIR (default .)" |
Denys Vlasenko | dff9fef | 2017-01-24 21:41:43 +0100 | [diff] [blame] | 144 | //usage: "\n -S Show server response" |
Denys Vlasenko | 2972e2c | 2016-10-04 04:23:09 +0200 | [diff] [blame] | 145 | //usage: IF_FEATURE_WGET_TIMEOUT( |
| 146 | //usage: "\n -T SEC Network read timeout is SEC seconds" |
| 147 | //usage: ) |
| 148 | //usage: "\n -O FILE Save to FILE ('-' for stdout)" |
| 149 | //usage: "\n -U STR Use STR for User-Agent header" |
| 150 | //usage: "\n -Y on/off Use proxy" |
Denys Vlasenko | e2e55b0 | 2011-03-21 00:37:05 +0100 | [diff] [blame] | 151 | |
Denis Vlasenko | b6adbf1 | 2007-05-26 19:00:18 +0000 | [diff] [blame] | 152 | #include "libbb.h" |
Denis Vlasenko | a552eeb | 2006-09-26 09:22:12 +0000 | [diff] [blame] | 153 | |
Vladimir Dronnikov | f5abc78 | 2012-06-13 17:29:41 +0200 | [diff] [blame] | 154 | #if 0 |
| 155 | # define log_io(...) bb_error_msg(__VA_ARGS__) |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 156 | # define SENDFMT(fp, fmt, ...) \ |
| 157 | do { \ |
| 158 | log_io("> " fmt, ##__VA_ARGS__); \ |
| 159 | fprintf(fp, fmt, ##__VA_ARGS__); \ |
| 160 | } while (0); |
Vladimir Dronnikov | f5abc78 | 2012-06-13 17:29:41 +0200 | [diff] [blame] | 161 | #else |
| 162 | # define log_io(...) ((void)0) |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 163 | # define SENDFMT(fp, fmt, ...) fprintf(fp, fmt, ##__VA_ARGS__) |
Vladimir Dronnikov | f5abc78 | 2012-06-13 17:29:41 +0200 | [diff] [blame] | 164 | #endif |
Denys Vlasenko | f836f01 | 2011-02-10 23:02:28 +0100 | [diff] [blame] | 165 | |
| 166 | |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 167 | #define SSL_SUPPORTED (ENABLE_FEATURE_WGET_OPENSSL || ENABLE_FEATURE_WGET_HTTPS) |
Denys Vlasenko | a6f8651 | 2017-01-11 20:16:45 +0100 | [diff] [blame] | 168 | |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 169 | struct host_info { |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 170 | char *allocated; |
Denis Vlasenko | 818322b | 2007-09-24 18:27:04 +0000 | [diff] [blame] | 171 | const char *path; |
Denys Vlasenko | d353bff | 2014-02-03 14:09:42 +0100 | [diff] [blame] | 172 | char *user; |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 173 | const char *protocol; |
Denis Vlasenko | 818322b | 2007-09-24 18:27:04 +0000 | [diff] [blame] | 174 | char *host; |
| 175 | int port; |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 176 | }; |
Denys Vlasenko | 3e134eb | 2016-04-22 18:09:21 +0200 | [diff] [blame] | 177 | static const char P_FTP[] ALIGN1 = "ftp"; |
| 178 | static const char P_HTTP[] ALIGN1 = "http"; |
Denys Vlasenko | a6f8651 | 2017-01-11 20:16:45 +0100 | [diff] [blame] | 179 | #if SSL_SUPPORTED |
Denys Vlasenko | 403f299 | 2018-02-06 15:15:08 +0100 | [diff] [blame] | 180 | # if ENABLE_FEATURE_WGET_HTTPS |
| 181 | static const char P_FTPS[] ALIGN1 = "ftps"; |
| 182 | # endif |
Denys Vlasenko | 3e134eb | 2016-04-22 18:09:21 +0200 | [diff] [blame] | 183 | static const char P_HTTPS[] ALIGN1 = "https"; |
Ron Yorston | 4d0c1ea | 2015-10-12 10:51:25 +0100 | [diff] [blame] | 184 | #endif |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 185 | |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 186 | #if ENABLE_FEATURE_WGET_LONG_OPTIONS |
| 187 | /* User-specified headers prevent using our corresponding built-in headers. */ |
| 188 | enum { |
| 189 | HDR_HOST = (1<<0), |
| 190 | HDR_USER_AGENT = (1<<1), |
| 191 | HDR_RANGE = (1<<2), |
| 192 | HDR_AUTH = (1<<3) * ENABLE_FEATURE_WGET_AUTHENTICATION, |
| 193 | HDR_PROXY_AUTH = (1<<4) * ENABLE_FEATURE_WGET_AUTHENTICATION, |
| 194 | }; |
| 195 | static const char wget_user_headers[] ALIGN1 = |
| 196 | "Host:\0" |
| 197 | "User-Agent:\0" |
| 198 | "Range:\0" |
| 199 | # if ENABLE_FEATURE_WGET_AUTHENTICATION |
| 200 | "Authorization:\0" |
| 201 | "Proxy-Authorization:\0" |
| 202 | # endif |
| 203 | ; |
| 204 | # define USR_HEADER_HOST (G.user_headers & HDR_HOST) |
| 205 | # define USR_HEADER_USER_AGENT (G.user_headers & HDR_USER_AGENT) |
| 206 | # define USR_HEADER_RANGE (G.user_headers & HDR_RANGE) |
| 207 | # define USR_HEADER_AUTH (G.user_headers & HDR_AUTH) |
| 208 | # define USR_HEADER_PROXY_AUTH (G.user_headers & HDR_PROXY_AUTH) |
| 209 | #else /* No long options, no user-headers :( */ |
| 210 | # define USR_HEADER_HOST 0 |
| 211 | # define USR_HEADER_USER_AGENT 0 |
| 212 | # define USR_HEADER_RANGE 0 |
| 213 | # define USR_HEADER_AUTH 0 |
| 214 | # define USR_HEADER_PROXY_AUTH 0 |
| 215 | #endif |
Denis Vlasenko | 7710563 | 2007-09-24 15:04:00 +0000 | [diff] [blame] | 216 | |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 217 | /* Globals */ |
Denis Vlasenko | 7710563 | 2007-09-24 15:04:00 +0000 | [diff] [blame] | 218 | struct globals { |
| 219 | off_t content_len; /* Content-length of the file */ |
| 220 | off_t beg_range; /* Range at which continue begins */ |
| 221 | #if ENABLE_FEATURE_WGET_STATUSBAR |
Denis Vlasenko | 7710563 | 2007-09-24 15:04:00 +0000 | [diff] [blame] | 222 | off_t transferred; /* Number of bytes transferred so far */ |
| 223 | const char *curfile; /* Name of current file being transferred */ |
Magnus Damm | f591499 | 2009-11-08 16:34:43 +0100 | [diff] [blame] | 224 | bb_progress_t pmt; |
Denis Vlasenko | 7710563 | 2007-09-24 15:04:00 +0000 | [diff] [blame] | 225 | #endif |
Denys Vlasenko | 982e87f | 2013-07-30 11:52:58 +0200 | [diff] [blame] | 226 | char *dir_prefix; |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 227 | #if ENABLE_FEATURE_WGET_LONG_OPTIONS |
Denys Vlasenko | 982e87f | 2013-07-30 11:52:58 +0200 | [diff] [blame] | 228 | char *post_data; |
| 229 | char *extra_headers; |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 230 | unsigned char user_headers; /* Headers mentioned by the user */ |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 231 | #endif |
Denys Vlasenko | 982e87f | 2013-07-30 11:52:58 +0200 | [diff] [blame] | 232 | char *fname_out; /* where to direct output (-O) */ |
| 233 | const char *proxy_flag; /* Use proxies if env vars are set */ |
| 234 | const char *user_agent; /* "User-Agent" header field */ |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 235 | #if ENABLE_FEATURE_WGET_TIMEOUT |
| 236 | unsigned timeout_seconds; |
Denys Vlasenko | 6701e91 | 2016-03-17 15:58:16 +0100 | [diff] [blame] | 237 | bool die_if_timed_out; |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 238 | #endif |
Denys Vlasenko | 2384a35 | 2011-02-15 00:58:36 +0100 | [diff] [blame] | 239 | int output_fd; |
| 240 | int o_flags; |
Denys Vlasenko | 7f43280 | 2009-06-28 01:02:24 +0200 | [diff] [blame] | 241 | smallint chunked; /* chunked transfer encoding */ |
| 242 | smallint got_clen; /* got content-length: from server */ |
Denys Vlasenko | df4e16c | 2011-02-10 06:29:06 +0100 | [diff] [blame] | 243 | /* Local downloads do benefit from big buffer. |
| 244 | * With 512 byte buffer, it was measured to be |
| 245 | * an order of magnitude slower than with big one. |
| 246 | */ |
| 247 | uint64_t just_to_align_next_member; |
| 248 | char wget_buf[CONFIG_FEATURE_COPYBUF_KB*1024]; |
Denys Vlasenko | 98a4c7c | 2010-02-04 15:00:15 +0100 | [diff] [blame] | 249 | } FIX_ALIASING; |
Denys Vlasenko | df4e16c | 2011-02-10 06:29:06 +0100 | [diff] [blame] | 250 | #define G (*ptr_to_globals) |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 251 | #define INIT_G() do { \ |
Denys Vlasenko | 982e87f | 2013-07-30 11:52:58 +0200 | [diff] [blame] | 252 | SET_PTR_TO_GLOBALS(xzalloc(sizeof(G))); \ |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 253 | } while (0) |
Guilherme Maciel Ferreira | 840ef17 | 2013-10-16 14:43:30 +0200 | [diff] [blame] | 254 | #define FINI_G() do { \ |
| 255 | FREE_PTR_TO_GLOBALS(); \ |
| 256 | } while (0) |
Denis Vlasenko | 7710563 | 2007-09-24 15:04:00 +0000 | [diff] [blame] | 257 | |
| 258 | |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 259 | /* Must match option string! */ |
| 260 | enum { |
| 261 | WGET_OPT_CONTINUE = (1 << 0), |
Denys Vlasenko | 2972e2c | 2016-10-04 04:23:09 +0200 | [diff] [blame] | 262 | WGET_OPT_QUIET = (1 << 1), |
Denys Vlasenko | dff9fef | 2017-01-24 21:41:43 +0100 | [diff] [blame] | 263 | WGET_OPT_SERVER_RESPONSE = (1 << 2), |
| 264 | WGET_OPT_OUTNAME = (1 << 3), |
| 265 | WGET_OPT_PREFIX = (1 << 4), |
| 266 | WGET_OPT_PROXY = (1 << 5), |
| 267 | WGET_OPT_USER_AGENT = (1 << 6), |
| 268 | WGET_OPT_NETWORK_READ_TIMEOUT = (1 << 7), |
| 269 | WGET_OPT_RETRIES = (1 << 8), |
| 270 | WGET_OPT_nsomething = (1 << 9), |
| 271 | WGET_OPT_HEADER = (1 << 10) * ENABLE_FEATURE_WGET_LONG_OPTIONS, |
| 272 | WGET_OPT_POST_DATA = (1 << 11) * ENABLE_FEATURE_WGET_LONG_OPTIONS, |
| 273 | WGET_OPT_SPIDER = (1 << 12) * ENABLE_FEATURE_WGET_LONG_OPTIONS, |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 274 | }; |
| 275 | |
| 276 | enum { |
| 277 | PROGRESS_START = -1, |
| 278 | PROGRESS_END = 0, |
| 279 | PROGRESS_BUMP = 1, |
| 280 | }; |
Denis Vlasenko | 9cade08 | 2006-11-21 10:43:02 +0000 | [diff] [blame] | 281 | #if ENABLE_FEATURE_WGET_STATUSBAR |
Denis Vlasenko | 00d8417 | 2008-11-24 07:34:42 +0000 | [diff] [blame] | 282 | static void progress_meter(int flag) |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 283 | { |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 284 | if (option_mask32 & WGET_OPT_QUIET) |
| 285 | return; |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 286 | |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 287 | if (flag == PROGRESS_START) |
Denys Vlasenko | d55e139 | 2011-02-11 18:56:13 +0100 | [diff] [blame] | 288 | bb_progress_init(&G.pmt, G.curfile); |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 289 | |
Denys Vlasenko | 2384a35 | 2011-02-15 00:58:36 +0100 | [diff] [blame] | 290 | bb_progress_update(&G.pmt, |
| 291 | G.beg_range, |
| 292 | G.transferred, |
| 293 | (G.chunked || !G.got_clen) ? 0 : G.beg_range + G.transferred + G.content_len |
| 294 | ); |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 295 | |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 296 | if (flag == PROGRESS_END) { |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 297 | bb_progress_free(&G.pmt); |
Denys Vlasenko | 19ced5c | 2010-06-06 21:53:09 +0200 | [diff] [blame] | 298 | bb_putchar_stderr('\n'); |
Denys Vlasenko | a3aa3e3 | 2009-12-11 12:36:10 +0100 | [diff] [blame] | 299 | G.transferred = 0; |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 300 | } |
| 301 | } |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 302 | #else |
Denis Vlasenko | 00d8417 | 2008-11-24 07:34:42 +0000 | [diff] [blame] | 303 | static ALWAYS_INLINE void progress_meter(int flag UNUSED_PARAM) { } |
Eric Andersen | b520e08 | 2000-10-03 00:21:45 +0000 | [diff] [blame] | 304 | #endif |
Eric Andersen | c7bda1c | 2004-03-15 08:29:22 +0000 | [diff] [blame] | 305 | |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 306 | |
Denys Vlasenko | 7d5ddf1 | 2009-06-30 20:36:27 +0200 | [diff] [blame] | 307 | /* IPv6 knows scoped address types i.e. link and site local addresses. Link |
| 308 | * local addresses can have a scope identifier to specify the |
| 309 | * interface/link an address is valid on (e.g. fe80::1%eth0). This scope |
| 310 | * identifier is only valid on a single node. |
| 311 | * |
| 312 | * RFC 4007 says that the scope identifier MUST NOT be sent across the wire, |
| 313 | * unless all nodes agree on the semantic. Apache e.g. regards zone identifiers |
| 314 | * in the Host header as invalid requests, see |
| 315 | * https://issues.apache.org/bugzilla/show_bug.cgi?id=35122 |
| 316 | */ |
| 317 | static void strip_ipv6_scope_id(char *host) |
| 318 | { |
| 319 | char *scope, *cp; |
| 320 | |
| 321 | /* bbox wget actually handles IPv6 addresses without [], like |
| 322 | * wget "http://::1/xxx", but this is not standard. |
| 323 | * To save code, _here_ we do not support it. */ |
| 324 | |
| 325 | if (host[0] != '[') |
| 326 | return; /* not IPv6 */ |
| 327 | |
| 328 | scope = strchr(host, '%'); |
| 329 | if (!scope) |
| 330 | return; |
| 331 | |
| 332 | /* Remove the IPv6 zone identifier from the host address */ |
| 333 | cp = strchr(host, ']'); |
| 334 | if (!cp || (cp[1] != ':' && cp[1] != '\0')) { |
| 335 | /* malformed address (not "[xx]:nn" or "[xx]") */ |
| 336 | return; |
| 337 | } |
| 338 | |
| 339 | /* cp points to "]...", scope points to "%eth0]..." */ |
| 340 | overlapping_strcpy(scope, cp); |
| 341 | } |
| 342 | |
Denis Vlasenko | 9cade08 | 2006-11-21 10:43:02 +0000 | [diff] [blame] | 343 | #if ENABLE_FEATURE_WGET_AUTHENTICATION |
Denys Vlasenko | df4e16c | 2011-02-10 06:29:06 +0100 | [diff] [blame] | 344 | /* Base64-encode character string. */ |
| 345 | static char *base64enc(const char *str) |
Denis Vlasenko | 3526a13 | 2006-09-09 12:20:57 +0000 | [diff] [blame] | 346 | { |
Denis Vlasenko | 12d2129 | 2007-06-27 21:40:07 +0000 | [diff] [blame] | 347 | unsigned len = strlen(str); |
Denys Vlasenko | df4e16c | 2011-02-10 06:29:06 +0100 | [diff] [blame] | 348 | if (len > sizeof(G.wget_buf)/4*3 - 10) /* paranoia */ |
| 349 | len = sizeof(G.wget_buf)/4*3 - 10; |
| 350 | bb_uuencode(G.wget_buf, str, len, bb_uuenc_tbl_base64); |
| 351 | return G.wget_buf; |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 352 | } |
| 353 | #endif |
| 354 | |
Lauri Kasanen | d074b41 | 2013-10-12 21:47:07 +0200 | [diff] [blame] | 355 | #if ENABLE_FEATURE_WGET_TIMEOUT |
| 356 | static void alarm_handler(int sig UNUSED_PARAM) |
| 357 | { |
| 358 | /* This is theoretically unsafe (uses stdio and malloc in signal handler) */ |
Denys Vlasenko | 6701e91 | 2016-03-17 15:58:16 +0100 | [diff] [blame] | 359 | if (G.die_if_timed_out) |
Lauri Kasanen | d074b41 | 2013-10-12 21:47:07 +0200 | [diff] [blame] | 360 | bb_error_msg_and_die("download timed out"); |
| 361 | } |
Denys Vlasenko | 6701e91 | 2016-03-17 15:58:16 +0100 | [diff] [blame] | 362 | static void set_alarm(void) |
| 363 | { |
| 364 | if (G.timeout_seconds) { |
| 365 | alarm(G.timeout_seconds); |
| 366 | G.die_if_timed_out = 1; |
| 367 | } |
| 368 | } |
| 369 | # define clear_alarm() ((void)(G.die_if_timed_out = 0)) |
| 370 | #else |
| 371 | # define set_alarm() ((void)0) |
| 372 | # define clear_alarm() ((void)0) |
Lauri Kasanen | d074b41 | 2013-10-12 21:47:07 +0200 | [diff] [blame] | 373 | #endif |
| 374 | |
Denys Vlasenko | ed72761 | 2016-07-25 21:34:57 +0200 | [diff] [blame] | 375 | #if ENABLE_FEATURE_WGET_OPENSSL |
| 376 | /* |
| 377 | * is_ip_address() attempts to verify whether or not a string |
| 378 | * contains an IPv4 or IPv6 address (vs. an FQDN). The result |
| 379 | * of inet_pton() can be used to determine this. |
| 380 | * |
| 381 | * TODO add proper error checking when inet_pton() returns -1 |
| 382 | * (some form of system error has occurred, and errno is set) |
| 383 | */ |
| 384 | static int is_ip_address(const char *string) |
| 385 | { |
| 386 | struct sockaddr_in sa; |
| 387 | |
| 388 | int result = inet_pton(AF_INET, string, &(sa.sin_addr)); |
| 389 | # if ENABLE_FEATURE_IPV6 |
| 390 | if (result == 0) { |
| 391 | struct sockaddr_in6 sa6; |
| 392 | result = inet_pton(AF_INET6, string, &(sa6.sin6_addr)); |
| 393 | } |
| 394 | # endif |
| 395 | return (result == 1); |
| 396 | } |
| 397 | #endif |
| 398 | |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 399 | static FILE *open_socket(len_and_sockaddr *lsa) |
| 400 | { |
Lauri Kasanen | d074b41 | 2013-10-12 21:47:07 +0200 | [diff] [blame] | 401 | int fd; |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 402 | FILE *fp; |
| 403 | |
Denys Vlasenko | 6701e91 | 2016-03-17 15:58:16 +0100 | [diff] [blame] | 404 | set_alarm(); |
Lauri Kasanen | d074b41 | 2013-10-12 21:47:07 +0200 | [diff] [blame] | 405 | fd = xconnect_stream(lsa); |
Denys Vlasenko | 6701e91 | 2016-03-17 15:58:16 +0100 | [diff] [blame] | 406 | clear_alarm(); |
Lauri Kasanen | d074b41 | 2013-10-12 21:47:07 +0200 | [diff] [blame] | 407 | |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 408 | /* glibc 2.4 seems to try seeking on it - ??! */ |
| 409 | /* hopefully it understands what ESPIPE means... */ |
Lauri Kasanen | d074b41 | 2013-10-12 21:47:07 +0200 | [diff] [blame] | 410 | fp = fdopen(fd, "r+"); |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 411 | if (!fp) |
Denys Vlasenko | dee0fc9 | 2011-02-10 10:01:49 +0100 | [diff] [blame] | 412 | bb_perror_msg_and_die(bb_msg_memory_exhausted); |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 413 | |
| 414 | return fp; |
| 415 | } |
| 416 | |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 417 | /* We balk at any control chars in other side's messages. |
| 418 | * This prevents nasty surprises (e.g. ESC sequences) in "Location:" URLs |
| 419 | * and error messages. |
| 420 | * |
| 421 | * The only exception is tabs, which are converted to (one) space: |
| 422 | * HTTP's "headers: <whitespace> values" may have those. |
| 423 | */ |
| 424 | static char* sanitize_string(char *s) |
| 425 | { |
| 426 | unsigned char *p = (void *) s; |
| 427 | while (*p) { |
| 428 | if (*p < ' ') { |
| 429 | if (*p != '\t') |
| 430 | break; |
| 431 | *p = ' '; |
| 432 | } |
| 433 | p++; |
| 434 | } |
| 435 | *p = '\0'; |
| 436 | return s; |
| 437 | } |
| 438 | |
Denys Vlasenko | f836f01 | 2011-02-10 23:02:28 +0100 | [diff] [blame] | 439 | /* Returns '\n' if it was seen, else '\0'. Trims at first '\r' or '\n' */ |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 440 | static char fgets_trim_sanitize(FILE *fp, const char *fmt) |
Denys Vlasenko | f836f01 | 2011-02-10 23:02:28 +0100 | [diff] [blame] | 441 | { |
| 442 | char c; |
| 443 | char *buf_ptr; |
| 444 | |
Denys Vlasenko | 6701e91 | 2016-03-17 15:58:16 +0100 | [diff] [blame] | 445 | set_alarm(); |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 446 | if (fgets(G.wget_buf, sizeof(G.wget_buf), fp) == NULL) |
Denys Vlasenko | f836f01 | 2011-02-10 23:02:28 +0100 | [diff] [blame] | 447 | bb_perror_msg_and_die("error getting response"); |
Denys Vlasenko | 6701e91 | 2016-03-17 15:58:16 +0100 | [diff] [blame] | 448 | clear_alarm(); |
Denys Vlasenko | f836f01 | 2011-02-10 23:02:28 +0100 | [diff] [blame] | 449 | |
| 450 | buf_ptr = strchrnul(G.wget_buf, '\n'); |
| 451 | c = *buf_ptr; |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 452 | #if 1 |
| 453 | /* Disallow any control chars: trim at first char < 0x20 */ |
| 454 | sanitize_string(G.wget_buf); |
| 455 | #else |
Denys Vlasenko | f836f01 | 2011-02-10 23:02:28 +0100 | [diff] [blame] | 456 | *buf_ptr = '\0'; |
| 457 | buf_ptr = strchrnul(G.wget_buf, '\r'); |
| 458 | *buf_ptr = '\0'; |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 459 | #endif |
Denys Vlasenko | f836f01 | 2011-02-10 23:02:28 +0100 | [diff] [blame] | 460 | |
| 461 | log_io("< %s", G.wget_buf); |
| 462 | |
Denys Vlasenko | dff9fef | 2017-01-24 21:41:43 +0100 | [diff] [blame] | 463 | if (fmt && (option_mask32 & WGET_OPT_SERVER_RESPONSE)) |
| 464 | fprintf(stderr, fmt, G.wget_buf); |
| 465 | |
Denys Vlasenko | f836f01 | 2011-02-10 23:02:28 +0100 | [diff] [blame] | 466 | return c; |
| 467 | } |
| 468 | |
Denys Vlasenko | df4e16c | 2011-02-10 06:29:06 +0100 | [diff] [blame] | 469 | static int ftpcmd(const char *s1, const char *s2, FILE *fp) |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 470 | { |
| 471 | int result; |
| 472 | if (s1) { |
Denys Vlasenko | f836f01 | 2011-02-10 23:02:28 +0100 | [diff] [blame] | 473 | if (!s2) |
| 474 | s2 = ""; |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 475 | fprintf(fp, "%s%s\r\n", s1, s2); |
Denys Vlasenko | dff9fef | 2017-01-24 21:41:43 +0100 | [diff] [blame] | 476 | /* With --server-response, wget also shows its ftp commands */ |
| 477 | if (option_mask32 & WGET_OPT_SERVER_RESPONSE) |
| 478 | fprintf(stderr, "--> %s%s\n\n", s1, s2); |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 479 | fflush(fp); |
Denys Vlasenko | f836f01 | 2011-02-10 23:02:28 +0100 | [diff] [blame] | 480 | log_io("> %s%s", s1, s2); |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 481 | } |
| 482 | |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 483 | /* Read until "Nxx something" is received */ |
| 484 | G.wget_buf[3] = 0; |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 485 | do { |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 486 | fgets_trim_sanitize(fp, "%s\n"); |
Denys Vlasenko | df4e16c | 2011-02-10 06:29:06 +0100 | [diff] [blame] | 487 | } while (!isdigit(G.wget_buf[0]) || G.wget_buf[3] != ' '); |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 488 | |
Denys Vlasenko | df4e16c | 2011-02-10 06:29:06 +0100 | [diff] [blame] | 489 | G.wget_buf[3] = '\0'; |
| 490 | result = xatoi_positive(G.wget_buf); |
| 491 | G.wget_buf[3] = ' '; |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 492 | return result; |
| 493 | } |
| 494 | |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 495 | static void parse_url(const char *src_url, struct host_info *h) |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 496 | { |
| 497 | char *url, *p, *sp; |
| 498 | |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 499 | free(h->allocated); |
| 500 | h->allocated = url = xstrdup(src_url); |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 501 | |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 502 | h->protocol = P_FTP; |
| 503 | p = strstr(url, "://"); |
| 504 | if (p) { |
| 505 | *p = '\0'; |
| 506 | h->host = p + 3; |
| 507 | if (strcmp(url, P_FTP) == 0) { |
| 508 | h->port = bb_lookup_port(P_FTP, "tcp", 21); |
| 509 | } else |
Denys Vlasenko | a6f8651 | 2017-01-11 20:16:45 +0100 | [diff] [blame] | 510 | #if SSL_SUPPORTED |
Denys Vlasenko | 403f299 | 2018-02-06 15:15:08 +0100 | [diff] [blame] | 511 | # if ENABLE_FEATURE_WGET_HTTPS |
| 512 | if (strcmp(url, P_FTPS) == 0) { |
| 513 | h->port = bb_lookup_port(P_FTPS, "tcp", 990); |
| 514 | h->protocol = P_FTPS; |
| 515 | } else |
| 516 | # endif |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 517 | if (strcmp(url, P_HTTPS) == 0) { |
| 518 | h->port = bb_lookup_port(P_HTTPS, "tcp", 443); |
| 519 | h->protocol = P_HTTPS; |
| 520 | } else |
Ron Yorston | 4d0c1ea | 2015-10-12 10:51:25 +0100 | [diff] [blame] | 521 | #endif |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 522 | if (strcmp(url, P_HTTP) == 0) { |
Lauri Kasanen | 4967a41 | 2013-12-17 19:03:41 +0100 | [diff] [blame] | 523 | http: |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 524 | h->port = bb_lookup_port(P_HTTP, "tcp", 80); |
| 525 | h->protocol = P_HTTP; |
| 526 | } else { |
| 527 | *p = ':'; |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 528 | bb_error_msg_and_die("not an http or ftp url: %s", url); |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 529 | } |
| 530 | } else { |
Lauri Kasanen | 4967a41 | 2013-12-17 19:03:41 +0100 | [diff] [blame] | 531 | // GNU wget is user-friendly and falls back to http:// |
| 532 | h->host = url; |
| 533 | goto http; |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 534 | } |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 535 | |
| 536 | // FYI: |
| 537 | // "Real" wget 'http://busybox.net?var=a/b' sends this request: |
Denys Vlasenko | a0aae9f | 2017-01-20 14:12:10 +0100 | [diff] [blame] | 538 | // 'GET /?var=a/b HTTP/1.0' |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 539 | // and saves 'index.html?var=a%2Fb' (we save 'b') |
| 540 | // wget 'http://busybox.net?login=john@doe': |
| 541 | // request: 'GET /?login=john@doe HTTP/1.0' |
| 542 | // saves: 'index.html?login=john@doe' (we save '?login=john@doe') |
| 543 | // wget 'http://busybox.net#test/test': |
| 544 | // request: 'GET / HTTP/1.0' |
| 545 | // saves: 'index.html' (we save 'test') |
| 546 | // |
| 547 | // We also don't add unique .N suffix if file exists... |
| 548 | sp = strchr(h->host, '/'); |
| 549 | p = strchr(h->host, '?'); if (!sp || (p && sp > p)) sp = p; |
| 550 | p = strchr(h->host, '#'); if (!sp || (p && sp > p)) sp = p; |
| 551 | if (!sp) { |
Denis Vlasenko | 818322b | 2007-09-24 18:27:04 +0000 | [diff] [blame] | 552 | h->path = ""; |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 553 | } else if (*sp == '/') { |
| 554 | *sp = '\0'; |
| 555 | h->path = sp + 1; |
| 556 | } else { // '#' or '?' |
| 557 | // http://busybox.net?login=john@doe is a valid URL |
| 558 | // memmove converts to: |
| 559 | // http:/busybox.nett?login=john@doe... |
Denis Vlasenko | 818322b | 2007-09-24 18:27:04 +0000 | [diff] [blame] | 560 | memmove(h->host - 1, h->host, sp - h->host); |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 561 | h->host--; |
| 562 | sp[-1] = '\0'; |
| 563 | h->path = sp; |
| 564 | } |
| 565 | |
| 566 | sp = strrchr(h->host, '@'); |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 567 | if (sp != NULL) { |
Denys Vlasenko | dd1061b | 2011-09-11 21:04:02 +0200 | [diff] [blame] | 568 | // URL-decode "user:password" string before base64-encoding: |
| 569 | // wget http://test:my%20pass@example.com should send |
| 570 | // Authorization: Basic dGVzdDpteSBwYXNz |
| 571 | // which decodes to "test:my pass". |
| 572 | // Standard wget and curl do this too. |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 573 | *sp = '\0'; |
Denys Vlasenko | d353bff | 2014-02-03 14:09:42 +0100 | [diff] [blame] | 574 | free(h->user); |
| 575 | h->user = xstrdup(percent_decode_in_place(h->host, /*strict:*/ 0)); |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 576 | h->host = sp + 1; |
| 577 | } |
Denys Vlasenko | d353bff | 2014-02-03 14:09:42 +0100 | [diff] [blame] | 578 | /* else: h->user remains NULL, or as set by original request |
| 579 | * before redirect (if we are here after a redirect). |
| 580 | */ |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 581 | } |
| 582 | |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 583 | static char *get_sanitized_hdr(FILE *fp) |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 584 | { |
| 585 | char *s, *hdrval; |
| 586 | int c; |
| 587 | |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 588 | /* retrieve header line */ |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 589 | c = fgets_trim_sanitize(fp, " %s\n"); |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 590 | |
Denys Vlasenko | f836f01 | 2011-02-10 23:02:28 +0100 | [diff] [blame] | 591 | /* end of the headers? */ |
| 592 | if (G.wget_buf[0] == '\0') |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 593 | return NULL; |
| 594 | |
| 595 | /* convert the header name to lower case */ |
Denys Vlasenko | ea267d5 | 2013-07-01 15:01:50 +0200 | [diff] [blame] | 596 | for (s = G.wget_buf; isalnum(*s) || *s == '-' || *s == '.' || *s == '_'; ++s) { |
| 597 | /* |
| 598 | * No-op for 20-3f and 60-7f. "0-9a-z-." are in these ranges. |
| 599 | * 40-5f range ("@A-Z[\]^_") maps to 60-7f. |
| 600 | * "A-Z" maps to "a-z". |
| 601 | * "@[\]" can't occur in header names. |
| 602 | * "^_" maps to "~,DEL" (which is wrong). |
| 603 | * "^" was never seen yet, "_" was seen from web.archive.org |
| 604 | * (x-archive-orig-x_commoncrawl_Signature: HEXSTRING). |
| 605 | */ |
Denys Vlasenko | f836f01 | 2011-02-10 23:02:28 +0100 | [diff] [blame] | 606 | *s |= 0x20; |
Denys Vlasenko | 4836331 | 2010-04-04 15:29:32 +0200 | [diff] [blame] | 607 | } |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 608 | |
| 609 | /* verify we are at the end of the header name */ |
| 610 | if (*s != ':') |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 611 | bb_error_msg_and_die("bad header line: %s", G.wget_buf); |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 612 | |
| 613 | /* locate the start of the header value */ |
| 614 | *s++ = '\0'; |
| 615 | hdrval = skip_whitespace(s); |
| 616 | |
Denys Vlasenko | f836f01 | 2011-02-10 23:02:28 +0100 | [diff] [blame] | 617 | if (c != '\n') { |
| 618 | /* Rats! The buffer isn't big enough to hold the entire header value */ |
| 619 | while (c = getc(fp), c != EOF && c != '\n') |
| 620 | continue; |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 621 | } |
| 622 | |
Denis Vlasenko | 47ddd01 | 2007-09-24 18:24:17 +0000 | [diff] [blame] | 623 | return hdrval; |
| 624 | } |
| 625 | |
Denys Vlasenko | bf146b8 | 2012-06-13 17:31:07 +0200 | [diff] [blame] | 626 | static void reset_beg_range_to_zero(void) |
| 627 | { |
Denys Vlasenko | 6144124 | 2012-06-17 19:52:25 +0200 | [diff] [blame] | 628 | bb_error_msg("restart failed"); |
Denys Vlasenko | bf146b8 | 2012-06-13 17:31:07 +0200 | [diff] [blame] | 629 | G.beg_range = 0; |
| 630 | xlseek(G.output_fd, 0, SEEK_SET); |
Denys Vlasenko | 6144124 | 2012-06-17 19:52:25 +0200 | [diff] [blame] | 631 | /* Done at the end instead: */ |
| 632 | /* ftruncate(G.output_fd, 0); */ |
Denys Vlasenko | bf146b8 | 2012-06-13 17:31:07 +0200 | [diff] [blame] | 633 | } |
| 634 | |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 635 | #if ENABLE_FEATURE_WGET_OPENSSL |
Denys Vlasenko | 2007ef5 | 2015-10-07 02:40:53 +0200 | [diff] [blame] | 636 | static int spawn_https_helper_openssl(const char *host, unsigned port) |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 637 | { |
| 638 | char *allocated = NULL; |
Denys Vlasenko | ed72761 | 2016-07-25 21:34:57 +0200 | [diff] [blame] | 639 | char *servername; |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 640 | int sp[2]; |
| 641 | int pid; |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 642 | IF_FEATURE_WGET_HTTPS(volatile int child_failed = 0;) |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 643 | |
| 644 | if (socketpair(AF_UNIX, SOCK_STREAM, 0, sp) != 0) |
| 645 | /* Kernel can have AF_UNIX support disabled */ |
| 646 | bb_perror_msg_and_die("socketpair"); |
| 647 | |
| 648 | if (!strchr(host, ':')) |
| 649 | host = allocated = xasprintf("%s:%u", host, port); |
Denys Vlasenko | ed72761 | 2016-07-25 21:34:57 +0200 | [diff] [blame] | 650 | servername = xstrdup(host); |
| 651 | strrchr(servername, ':')[0] = '\0'; |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 652 | |
Denys Vlasenko | 2007ef5 | 2015-10-07 02:40:53 +0200 | [diff] [blame] | 653 | fflush_all(); |
| 654 | pid = xvfork(); |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 655 | if (pid == 0) { |
| 656 | /* Child */ |
Denys Vlasenko | ed72761 | 2016-07-25 21:34:57 +0200 | [diff] [blame] | 657 | char *argv[8]; |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 658 | |
| 659 | close(sp[0]); |
| 660 | xmove_fd(sp[1], 0); |
| 661 | xdup2(0, 1); |
| 662 | /* |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 663 | * openssl s_client -quiet -connect www.kernel.org:443 2>/dev/null |
| 664 | * It prints some debug stuff on stderr, don't know how to suppress it. |
| 665 | * Work around by dev-nulling stderr. We lose all error messages :( |
| 666 | */ |
| 667 | xmove_fd(2, 3); |
| 668 | xopen("/dev/null", O_RDWR); |
Denys Vlasenko | ed72761 | 2016-07-25 21:34:57 +0200 | [diff] [blame] | 669 | memset(&argv, 0, sizeof(argv)); |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 670 | argv[0] = (char*)"openssl"; |
| 671 | argv[1] = (char*)"s_client"; |
| 672 | argv[2] = (char*)"-quiet"; |
| 673 | argv[3] = (char*)"-connect"; |
| 674 | argv[4] = (char*)host; |
Denys Vlasenko | ed72761 | 2016-07-25 21:34:57 +0200 | [diff] [blame] | 675 | /* |
| 676 | * Per RFC 6066 Section 3, the only permitted values in the |
| 677 | * TLS server_name (SNI) field are FQDNs (DNS hostnames). |
| 678 | * IPv4 and IPv6 addresses, port numbers are not allowed. |
| 679 | */ |
| 680 | if (!is_ip_address(servername)) { |
| 681 | argv[5] = (char*)"-servername"; |
| 682 | argv[6] = (char*)servername; |
| 683 | } |
| 684 | |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 685 | BB_EXECVP(argv[0], argv); |
| 686 | xmove_fd(3, 2); |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 687 | # if ENABLE_FEATURE_WGET_HTTPS |
Denys Vlasenko | 2007ef5 | 2015-10-07 02:40:53 +0200 | [diff] [blame] | 688 | child_failed = 1; |
| 689 | xfunc_die(); |
| 690 | # else |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 691 | bb_perror_msg_and_die("can't execute '%s'", argv[0]); |
Denys Vlasenko | 2007ef5 | 2015-10-07 02:40:53 +0200 | [diff] [blame] | 692 | # endif |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 693 | /* notreached */ |
| 694 | } |
| 695 | |
Denys Vlasenko | 5331557 | 2014-02-23 23:39:47 +0100 | [diff] [blame] | 696 | /* Parent */ |
Denys Vlasenko | ed72761 | 2016-07-25 21:34:57 +0200 | [diff] [blame] | 697 | free(servername); |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 698 | free(allocated); |
| 699 | close(sp[1]); |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 700 | # if ENABLE_FEATURE_WGET_HTTPS |
Denys Vlasenko | 2007ef5 | 2015-10-07 02:40:53 +0200 | [diff] [blame] | 701 | if (child_failed) { |
| 702 | close(sp[0]); |
| 703 | return -1; |
| 704 | } |
| 705 | # endif |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 706 | return sp[0]; |
| 707 | } |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 708 | #endif |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 709 | |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 710 | #if ENABLE_FEATURE_WGET_HTTPS |
Denys Vlasenko | 403f299 | 2018-02-06 15:15:08 +0100 | [diff] [blame] | 711 | static void spawn_ssl_client(const char *host, int network_fd, int flags) |
Denys Vlasenko | 5331557 | 2014-02-23 23:39:47 +0100 | [diff] [blame] | 712 | { |
| 713 | int sp[2]; |
| 714 | int pid; |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 715 | char *servername, *p; |
| 716 | |
| 717 | servername = xstrdup(host); |
| 718 | p = strrchr(servername, ':'); |
| 719 | if (p) *p = '\0'; |
Denys Vlasenko | 5331557 | 2014-02-23 23:39:47 +0100 | [diff] [blame] | 720 | |
| 721 | if (socketpair(AF_UNIX, SOCK_STREAM, 0, sp) != 0) |
| 722 | /* Kernel can have AF_UNIX support disabled */ |
| 723 | bb_perror_msg_and_die("socketpair"); |
| 724 | |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 725 | fflush_all(); |
Denys Vlasenko | 5331557 | 2014-02-23 23:39:47 +0100 | [diff] [blame] | 726 | pid = BB_MMU ? xfork() : xvfork(); |
| 727 | if (pid == 0) { |
| 728 | /* Child */ |
Denys Vlasenko | 5331557 | 2014-02-23 23:39:47 +0100 | [diff] [blame] | 729 | close(sp[0]); |
| 730 | xmove_fd(sp[1], 0); |
| 731 | xdup2(0, 1); |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 732 | if (BB_MMU) { |
| 733 | tls_state_t *tls = new_tls_state(); |
| 734 | tls->ifd = tls->ofd = network_fd; |
| 735 | tls_handshake(tls, servername); |
Denys Vlasenko | 403f299 | 2018-02-06 15:15:08 +0100 | [diff] [blame] | 736 | tls_run_copy_loop(tls, flags); |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 737 | exit(0); |
| 738 | } else { |
Denys Vlasenko | 403f299 | 2018-02-06 15:15:08 +0100 | [diff] [blame] | 739 | char *argv[6]; |
| 740 | |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 741 | xmove_fd(network_fd, 3); |
| 742 | argv[0] = (char*)"ssl_client"; |
| 743 | argv[1] = (char*)"-s3"; |
| 744 | //TODO: if (!is_ip_address(servername))... |
| 745 | argv[2] = (char*)"-n"; |
| 746 | argv[3] = servername; |
Denys Vlasenko | 403f299 | 2018-02-06 15:15:08 +0100 | [diff] [blame] | 747 | argv[4] = (flags & TLSLOOP_EXIT_ON_LOCAL_EOF ? (char*)"-e" : NULL); |
| 748 | argv[5] = NULL; |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 749 | BB_EXECVP(argv[0], argv); |
| 750 | bb_perror_msg_and_die("can't execute '%s'", argv[0]); |
| 751 | } |
Denys Vlasenko | 5331557 | 2014-02-23 23:39:47 +0100 | [diff] [blame] | 752 | /* notreached */ |
| 753 | } |
| 754 | |
| 755 | /* Parent */ |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 756 | free(servername); |
Denys Vlasenko | 5331557 | 2014-02-23 23:39:47 +0100 | [diff] [blame] | 757 | close(sp[1]); |
| 758 | xmove_fd(sp[0], network_fd); |
| 759 | } |
| 760 | #endif |
| 761 | |
Denys Vlasenko | e999657 | 2018-02-06 15:02:16 +0100 | [diff] [blame] | 762 | static FILE* prepare_ftp_session(FILE **dfpp, struct host_info *target, len_and_sockaddr *lsa) |
| 763 | { |
| 764 | FILE *sfp; |
| 765 | char *str; |
| 766 | int port; |
| 767 | |
| 768 | if (!target->user) |
| 769 | target->user = xstrdup("anonymous:busybox@"); |
| 770 | |
| 771 | sfp = open_socket(lsa); |
Denys Vlasenko | 403f299 | 2018-02-06 15:15:08 +0100 | [diff] [blame] | 772 | #if ENABLE_FEATURE_WGET_HTTPS |
| 773 | if (target->protocol == P_FTPS) |
| 774 | spawn_ssl_client(target->host, fileno(sfp), TLSLOOP_EXIT_ON_LOCAL_EOF); |
| 775 | #endif |
| 776 | |
Denys Vlasenko | e999657 | 2018-02-06 15:02:16 +0100 | [diff] [blame] | 777 | if (ftpcmd(NULL, NULL, sfp) != 220) |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 778 | bb_error_msg_and_die("%s", G.wget_buf); |
| 779 | /* note: ftpcmd() sanitizes G.wget_buf, ok to print */ |
Denys Vlasenko | e999657 | 2018-02-06 15:02:16 +0100 | [diff] [blame] | 780 | |
| 781 | /* |
| 782 | * Splitting username:password pair, |
| 783 | * trying to log in |
| 784 | */ |
| 785 | str = strchr(target->user, ':'); |
| 786 | if (str) |
| 787 | *str++ = '\0'; |
| 788 | switch (ftpcmd("USER ", target->user, sfp)) { |
| 789 | case 230: |
| 790 | break; |
| 791 | case 331: |
| 792 | if (ftpcmd("PASS ", str, sfp) == 230) |
| 793 | break; |
| 794 | /* fall through (failed login) */ |
| 795 | default: |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 796 | bb_error_msg_and_die("ftp login: %s", G.wget_buf); |
Denys Vlasenko | e999657 | 2018-02-06 15:02:16 +0100 | [diff] [blame] | 797 | } |
| 798 | |
| 799 | ftpcmd("TYPE I", NULL, sfp); |
| 800 | |
| 801 | /* |
| 802 | * Querying file size |
| 803 | */ |
| 804 | if (ftpcmd("SIZE ", target->path, sfp) == 213) { |
| 805 | G.content_len = BB_STRTOOFF(G.wget_buf + 4, NULL, 10); |
| 806 | if (G.content_len < 0 || errno) { |
| 807 | bb_error_msg_and_die("SIZE value is garbage"); |
| 808 | } |
| 809 | G.got_clen = 1; |
| 810 | } |
| 811 | |
| 812 | /* |
| 813 | * Entering passive mode |
| 814 | */ |
Denys Vlasenko | 1783ffa | 2018-02-06 15:48:12 +0100 | [diff] [blame] | 815 | if (ENABLE_FEATURE_IPV6 && ftpcmd("EPSV", NULL, sfp) == 229) { |
| 816 | /* good */ |
| 817 | } else |
Denys Vlasenko | e999657 | 2018-02-06 15:02:16 +0100 | [diff] [blame] | 818 | if (ftpcmd("PASV", NULL, sfp) != 227) { |
| 819 | pasv_error: |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 820 | bb_error_msg_and_die("bad response to %s: %s", "PASV", G.wget_buf); |
Denys Vlasenko | e999657 | 2018-02-06 15:02:16 +0100 | [diff] [blame] | 821 | } |
Denys Vlasenko | 1783ffa | 2018-02-06 15:48:12 +0100 | [diff] [blame] | 822 | port = parse_pasv_epsv(G.wget_buf); |
| 823 | if (port < 0) |
| 824 | goto pasv_error; |
| 825 | |
Denys Vlasenko | e999657 | 2018-02-06 15:02:16 +0100 | [diff] [blame] | 826 | set_nport(&lsa->u.sa, htons(port)); |
| 827 | |
| 828 | *dfpp = open_socket(lsa); |
| 829 | |
Denys Vlasenko | 2b75157 | 2018-02-06 20:49:27 +0100 | [diff] [blame] | 830 | #if ENABLE_FEATURE_WGET_HTTPS |
Denys Vlasenko | 237a900 | 2018-02-08 00:28:30 +0100 | [diff] [blame] | 831 | if (target->protocol == P_FTPS) { |
| 832 | /* "PROT P" enables encryption of data stream. |
| 833 | * Without it (or with "PROT C"), data is sent unencrypted. |
| 834 | */ |
| 835 | if (ftpcmd("PROT P", NULL, sfp) == 200) |
| 836 | spawn_ssl_client(target->host, fileno(*dfpp), /*flags*/ 0); |
| 837 | } |
Denys Vlasenko | 2b75157 | 2018-02-06 20:49:27 +0100 | [diff] [blame] | 838 | #endif |
Denys Vlasenko | 403f299 | 2018-02-06 15:15:08 +0100 | [diff] [blame] | 839 | |
Denys Vlasenko | e999657 | 2018-02-06 15:02:16 +0100 | [diff] [blame] | 840 | if (G.beg_range != 0) { |
| 841 | sprintf(G.wget_buf, "REST %"OFF_FMT"u", G.beg_range); |
| 842 | if (ftpcmd(G.wget_buf, NULL, sfp) == 350) |
| 843 | G.content_len -= G.beg_range; |
| 844 | else |
| 845 | reset_beg_range_to_zero(); |
| 846 | } |
| 847 | |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 848 | //TODO: needs ftp-escaping 0xff and '\n' bytes here. |
| 849 | //Or disallow '\n' altogether via sanitize_string() in parse_url(). |
| 850 | //But 0xff's are possible in valid utf8 filenames. |
Denys Vlasenko | e999657 | 2018-02-06 15:02:16 +0100 | [diff] [blame] | 851 | if (ftpcmd("RETR ", target->path, sfp) > 150) |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 852 | bb_error_msg_and_die("bad response to %s: %s", "RETR", G.wget_buf); |
Denys Vlasenko | e999657 | 2018-02-06 15:02:16 +0100 | [diff] [blame] | 853 | |
| 854 | return sfp; |
| 855 | } |
| 856 | |
Denys Vlasenko | 2384a35 | 2011-02-15 00:58:36 +0100 | [diff] [blame] | 857 | static void NOINLINE retrieve_file_data(FILE *dfp) |
Denys Vlasenko | 7f43280 | 2009-06-28 01:02:24 +0200 | [diff] [blame] | 858 | { |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 859 | #if ENABLE_FEATURE_WGET_STATUSBAR || ENABLE_FEATURE_WGET_TIMEOUT |
| 860 | # if ENABLE_FEATURE_WGET_TIMEOUT |
Denys Vlasenko | b7812ce | 2012-09-03 12:49:30 +0200 | [diff] [blame] | 861 | unsigned second_cnt = G.timeout_seconds; |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 862 | # endif |
| 863 | struct pollfd polldata; |
Denys Vlasenko | 7f43280 | 2009-06-28 01:02:24 +0200 | [diff] [blame] | 864 | |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 865 | polldata.fd = fileno(dfp); |
| 866 | polldata.events = POLLIN | POLLPRI; |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 867 | #endif |
| 868 | progress_meter(PROGRESS_START); |
Denys Vlasenko | 7f43280 | 2009-06-28 01:02:24 +0200 | [diff] [blame] | 869 | |
| 870 | if (G.chunked) |
| 871 | goto get_clen; |
| 872 | |
| 873 | /* Loops only if chunked */ |
| 874 | while (1) { |
Denys Vlasenko | c60f446 | 2011-02-11 22:23:23 +0100 | [diff] [blame] | 875 | |
| 876 | #if ENABLE_FEATURE_WGET_STATUSBAR || ENABLE_FEATURE_WGET_TIMEOUT |
| 877 | /* Must use nonblocking I/O, otherwise fread will loop |
| 878 | * and *block* until it reads full buffer, |
| 879 | * which messes up progress bar and/or timeout logic. |
| 880 | * Because of nonblocking I/O, we need to dance |
| 881 | * very carefully around EAGAIN. See explanation at |
Denys Vlasenko | b7812ce | 2012-09-03 12:49:30 +0200 | [diff] [blame] | 882 | * clearerr() calls. |
Denys Vlasenko | c60f446 | 2011-02-11 22:23:23 +0100 | [diff] [blame] | 883 | */ |
| 884 | ndelay_on(polldata.fd); |
| 885 | #endif |
Denys Vlasenko | a3aa3e3 | 2009-12-11 12:36:10 +0100 | [diff] [blame] | 886 | while (1) { |
Denys Vlasenko | 7f43280 | 2009-06-28 01:02:24 +0200 | [diff] [blame] | 887 | int n; |
Denys Vlasenko | a3aa3e3 | 2009-12-11 12:36:10 +0100 | [diff] [blame] | 888 | unsigned rdsz; |
Denys Vlasenko | 7f43280 | 2009-06-28 01:02:24 +0200 | [diff] [blame] | 889 | |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 890 | #if ENABLE_FEATURE_WGET_STATUSBAR || ENABLE_FEATURE_WGET_TIMEOUT |
Denys Vlasenko | 8766a79 | 2011-02-11 21:42:00 +0100 | [diff] [blame] | 891 | /* fread internally uses read loop, which in our case |
| 892 | * is usually exited when we get EAGAIN. |
| 893 | * In this case, libc sets error marker on the stream. |
| 894 | * Need to clear it before next fread to avoid possible |
| 895 | * rare false positive ferror below. Rare because usually |
| 896 | * fread gets more than zero bytes, and we don't fall |
| 897 | * into if (n <= 0) ... |
| 898 | */ |
| 899 | clearerr(dfp); |
Denys Vlasenko | f9af375 | 2011-02-11 22:01:33 +0100 | [diff] [blame] | 900 | #endif |
Denys Vlasenko | b7812ce | 2012-09-03 12:49:30 +0200 | [diff] [blame] | 901 | errno = 0; |
| 902 | rdsz = sizeof(G.wget_buf); |
| 903 | if (G.got_clen) { |
| 904 | if (G.content_len < (off_t)sizeof(G.wget_buf)) { |
| 905 | if ((int)G.content_len <= 0) |
| 906 | break; |
| 907 | rdsz = (unsigned)G.content_len; |
| 908 | } |
| 909 | } |
Denys Vlasenko | 0fac2f7 | 2011-02-10 09:55:05 +0100 | [diff] [blame] | 910 | n = fread(G.wget_buf, 1, rdsz, dfp); |
Denys Vlasenko | b7812ce | 2012-09-03 12:49:30 +0200 | [diff] [blame] | 911 | |
| 912 | if (n > 0) { |
| 913 | xwrite(G.output_fd, G.wget_buf, n); |
| 914 | #if ENABLE_FEATURE_WGET_STATUSBAR |
| 915 | G.transferred += n; |
| 916 | #endif |
| 917 | if (G.got_clen) { |
| 918 | G.content_len -= n; |
| 919 | if (G.content_len == 0) |
| 920 | break; |
| 921 | } |
| 922 | #if ENABLE_FEATURE_WGET_TIMEOUT |
| 923 | second_cnt = G.timeout_seconds; |
| 924 | #endif |
Denys Vlasenko | faa9e94 | 2014-03-27 16:50:29 +0100 | [diff] [blame] | 925 | goto bump; |
Denys Vlasenko | b7812ce | 2012-09-03 12:49:30 +0200 | [diff] [blame] | 926 | } |
| 927 | |
| 928 | /* n <= 0. |
| 929 | * man fread: |
Denys Vlasenko | 8766a79 | 2011-02-11 21:42:00 +0100 | [diff] [blame] | 930 | * If error occurs, or EOF is reached, the return value |
| 931 | * is a short item count (or zero). |
| 932 | * fread does not distinguish between EOF and error. |
| 933 | */ |
Denys Vlasenko | b7812ce | 2012-09-03 12:49:30 +0200 | [diff] [blame] | 934 | if (errno != EAGAIN) { |
| 935 | if (ferror(dfp)) { |
| 936 | progress_meter(PROGRESS_END); |
Denys Vlasenko | 8766a79 | 2011-02-11 21:42:00 +0100 | [diff] [blame] | 937 | bb_perror_msg_and_die(bb_msg_read_error); |
Denys Vlasenko | b7812ce | 2012-09-03 12:49:30 +0200 | [diff] [blame] | 938 | } |
Denys Vlasenko | 8766a79 | 2011-02-11 21:42:00 +0100 | [diff] [blame] | 939 | break; /* EOF, not error */ |
Denys Vlasenko | 7f43280 | 2009-06-28 01:02:24 +0200 | [diff] [blame] | 940 | } |
Denys Vlasenko | 8766a79 | 2011-02-11 21:42:00 +0100 | [diff] [blame] | 941 | |
Denys Vlasenko | b7812ce | 2012-09-03 12:49:30 +0200 | [diff] [blame] | 942 | #if ENABLE_FEATURE_WGET_STATUSBAR || ENABLE_FEATURE_WGET_TIMEOUT |
| 943 | /* It was EAGAIN. There is no data. Wait up to one second |
| 944 | * then abort if timed out, or update the bar and try reading again. |
| 945 | */ |
| 946 | if (safe_poll(&polldata, 1, 1000) == 0) { |
| 947 | # if ENABLE_FEATURE_WGET_TIMEOUT |
| 948 | if (second_cnt != 0 && --second_cnt == 0) { |
| 949 | progress_meter(PROGRESS_END); |
| 950 | bb_error_msg_and_die("download timed out"); |
| 951 | } |
| 952 | # endif |
| 953 | /* We used to loop back to poll here, |
| 954 | * but there is no great harm in letting fread |
| 955 | * to try reading anyway. |
| 956 | */ |
| 957 | } |
Denys Vlasenko | faa9e94 | 2014-03-27 16:50:29 +0100 | [diff] [blame] | 958 | #endif |
| 959 | bump: |
Denys Vlasenko | b7812ce | 2012-09-03 12:49:30 +0200 | [diff] [blame] | 960 | /* Need to do it _every_ second for "stalled" indicator |
| 961 | * to be shown properly. |
| 962 | */ |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 963 | progress_meter(PROGRESS_BUMP); |
Denys Vlasenko | b7812ce | 2012-09-03 12:49:30 +0200 | [diff] [blame] | 964 | } /* while (reading data) */ |
| 965 | |
Denys Vlasenko | c60f446 | 2011-02-11 22:23:23 +0100 | [diff] [blame] | 966 | #if ENABLE_FEATURE_WGET_STATUSBAR || ENABLE_FEATURE_WGET_TIMEOUT |
| 967 | clearerr(dfp); |
Denys Vlasenko | 88ad9da | 2011-02-11 23:06:21 +0100 | [diff] [blame] | 968 | ndelay_off(polldata.fd); /* else fgets can get very unhappy */ |
Denys Vlasenko | c60f446 | 2011-02-11 22:23:23 +0100 | [diff] [blame] | 969 | #endif |
Denys Vlasenko | 7f43280 | 2009-06-28 01:02:24 +0200 | [diff] [blame] | 970 | if (!G.chunked) |
| 971 | break; |
| 972 | |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 973 | fgets_trim_sanitize(dfp, NULL); /* Eat empty line */ |
Denys Vlasenko | 7f43280 | 2009-06-28 01:02:24 +0200 | [diff] [blame] | 974 | get_clen: |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 975 | fgets_trim_sanitize(dfp, NULL); |
Denys Vlasenko | df4e16c | 2011-02-10 06:29:06 +0100 | [diff] [blame] | 976 | G.content_len = STRTOOFF(G.wget_buf, NULL, 16); |
Denys Vlasenko | 7f43280 | 2009-06-28 01:02:24 +0200 | [diff] [blame] | 977 | /* FIXME: error check? */ |
Denys Vlasenko | a3aa3e3 | 2009-12-11 12:36:10 +0100 | [diff] [blame] | 978 | if (G.content_len == 0) |
Denys Vlasenko | 7f43280 | 2009-06-28 01:02:24 +0200 | [diff] [blame] | 979 | break; /* all done! */ |
Denys Vlasenko | a3aa3e3 | 2009-12-11 12:36:10 +0100 | [diff] [blame] | 980 | G.got_clen = 1; |
Denys Vlasenko | b7812ce | 2012-09-03 12:49:30 +0200 | [diff] [blame] | 981 | /* |
| 982 | * Note that fgets may result in some data being buffered in dfp. |
| 983 | * We loop back to fread, which will retrieve this data. |
| 984 | * Also note that code has to be arranged so that fread |
| 985 | * is done _before_ one-second poll wait - poll doesn't know |
| 986 | * about stdio buffering and can result in spurious one second waits! |
| 987 | */ |
Denys Vlasenko | 7f43280 | 2009-06-28 01:02:24 +0200 | [diff] [blame] | 988 | } |
| 989 | |
Denys Vlasenko | 6144124 | 2012-06-17 19:52:25 +0200 | [diff] [blame] | 990 | /* If -c failed, we restart from the beginning, |
| 991 | * but we do not truncate file then, we do it only now, at the end. |
| 992 | * This lets user to ^C if his 99% complete 10 GB file download |
| 993 | * failed to restart *without* losing the almost complete file. |
| 994 | */ |
| 995 | { |
| 996 | off_t pos = lseek(G.output_fd, 0, SEEK_CUR); |
| 997 | if (pos != (off_t)-1) |
| 998 | ftruncate(G.output_fd, pos); |
| 999 | } |
| 1000 | |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1001 | /* Draw full bar and free its resources */ |
Denys Vlasenko | 2384a35 | 2011-02-15 00:58:36 +0100 | [diff] [blame] | 1002 | G.chunked = 0; /* makes it show 100% even for chunked download */ |
| 1003 | G.got_clen = 1; /* makes it show 100% even for download of (formerly) unknown size */ |
Bradley M. Kuhn | c97131c | 2010-08-08 02:51:20 +0200 | [diff] [blame] | 1004 | progress_meter(PROGRESS_END); |
Denys Vlasenko | 7f43280 | 2009-06-28 01:02:24 +0200 | [diff] [blame] | 1005 | } |
| 1006 | |
Pere Orga | 5369563 | 2011-02-16 20:09:36 +0100 | [diff] [blame] | 1007 | static void download_one_url(const char *url) |
Eric Andersen | 9670083 | 2000-09-04 15:15:55 +0000 | [diff] [blame] | 1008 | { |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1009 | bool use_proxy; /* Use proxies if env vars are set */ |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1010 | int redir_limit; |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1011 | len_and_sockaddr *lsa; |
Denys Vlasenko | 7f43280 | 2009-06-28 01:02:24 +0200 | [diff] [blame] | 1012 | FILE *sfp; /* socket to web/ftp server */ |
Denis Vlasenko | a36535b | 2007-09-27 15:07:23 +0000 | [diff] [blame] | 1013 | FILE *dfp; /* socket to ftp server (data) */ |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1014 | char *fname_out_alloc; |
Denys Vlasenko | 93b4a60 | 2011-12-18 05:11:56 +0100 | [diff] [blame] | 1015 | char *redirected_path = NULL; |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1016 | struct host_info server; |
| 1017 | struct host_info target; |
Denis Vlasenko | 7710563 | 2007-09-24 15:04:00 +0000 | [diff] [blame] | 1018 | |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1019 | server.allocated = NULL; |
| 1020 | target.allocated = NULL; |
| 1021 | server.user = NULL; |
Vladimir Dronnikov | be168b1 | 2009-10-05 02:18:01 +0200 | [diff] [blame] | 1022 | target.user = NULL; |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1023 | |
| 1024 | parse_url(url, &target); |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 1025 | |
Bernhard Reutner-Fischer | 7e8a53a | 2007-04-10 09:37:29 +0000 | [diff] [blame] | 1026 | /* Use the proxy if necessary */ |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1027 | use_proxy = (strcmp(G.proxy_flag, "off") != 0); |
Glenn L McGrath | f1c4b11 | 2004-02-22 00:27:34 +0000 | [diff] [blame] | 1028 | if (use_proxy) { |
Peter Lloyd | 804ce5a | 2018-03-05 00:17:02 +0100 | [diff] [blame] | 1029 | char *proxy = getenv(target.protocol[0] == 'f' ? "ftp_proxy" : "http_proxy"); |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 1030 | //FIXME: what if protocol is https? Ok to use http_proxy? |
Denys Vlasenko | 2384a35 | 2011-02-15 00:58:36 +0100 | [diff] [blame] | 1031 | use_proxy = (proxy && proxy[0]); |
| 1032 | if (use_proxy) |
Denis Vlasenko | 96e9d3c | 2006-10-07 14:28:55 +0000 | [diff] [blame] | 1033 | parse_url(proxy, &server); |
Robert Griebl | d776011 | 2002-05-14 23:36:45 +0000 | [diff] [blame] | 1034 | } |
Denys Vlasenko | 7d5ddf1 | 2009-06-30 20:36:27 +0200 | [diff] [blame] | 1035 | if (!use_proxy) { |
Peter Lloyd | 804ce5a | 2018-03-05 00:17:02 +0100 | [diff] [blame] | 1036 | server.protocol = target.protocol; |
Denys Vlasenko | 7d5ddf1 | 2009-06-30 20:36:27 +0200 | [diff] [blame] | 1037 | server.port = target.port; |
| 1038 | if (ENABLE_FEATURE_IPV6) { |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1039 | //free(server.allocated); - can't be non-NULL |
| 1040 | server.host = server.allocated = xstrdup(target.host); |
Denys Vlasenko | 7d5ddf1 | 2009-06-30 20:36:27 +0200 | [diff] [blame] | 1041 | } else { |
| 1042 | server.host = target.host; |
| 1043 | } |
| 1044 | } |
| 1045 | |
| 1046 | if (ENABLE_FEATURE_IPV6) |
| 1047 | strip_ipv6_scope_id(target.host); |
Eric Andersen | c7bda1c | 2004-03-15 08:29:22 +0000 | [diff] [blame] | 1048 | |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1049 | /* If there was no -O FILE, guess output filename */ |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1050 | fname_out_alloc = NULL; |
Denys Vlasenko | 9a5b7f6 | 2011-02-13 02:49:43 +0100 | [diff] [blame] | 1051 | if (!(option_mask32 & WGET_OPT_OUTNAME)) { |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1052 | G.fname_out = bb_get_last_path_component_nostrip(target.path); |
Denis Vlasenko | 818322b | 2007-09-24 18:27:04 +0000 | [diff] [blame] | 1053 | /* handle "wget http://kernel.org//" */ |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1054 | if (G.fname_out[0] == '/' || !G.fname_out[0]) |
| 1055 | G.fname_out = (char*)"index.html"; |
Denis Vlasenko | 818322b | 2007-09-24 18:27:04 +0000 | [diff] [blame] | 1056 | /* -P DIR is considered only if there was no -O FILE */ |
Denys Vlasenko | aacd448 | 2012-06-17 20:21:30 +0200 | [diff] [blame] | 1057 | if (G.dir_prefix) |
| 1058 | G.fname_out = fname_out_alloc = concat_path_file(G.dir_prefix, G.fname_out); |
Denys Vlasenko | 625f218 | 2011-03-21 00:29:37 +0100 | [diff] [blame] | 1059 | else { |
Denys Vlasenko | aacd448 | 2012-06-17 20:21:30 +0200 | [diff] [blame] | 1060 | /* redirects may free target.path later, need to make a copy */ |
| 1061 | G.fname_out = fname_out_alloc = xstrdup(G.fname_out); |
Denys Vlasenko | 625f218 | 2011-03-21 00:29:37 +0100 | [diff] [blame] | 1062 | } |
Eric Andersen | 29edd00 | 2000-12-09 16:55:35 +0000 | [diff] [blame] | 1063 | } |
Denis Vlasenko | 818322b | 2007-09-24 18:27:04 +0000 | [diff] [blame] | 1064 | #if ENABLE_FEATURE_WGET_STATUSBAR |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1065 | G.curfile = bb_get_last_path_component_nostrip(G.fname_out); |
Denis Vlasenko | 818322b | 2007-09-24 18:27:04 +0000 | [diff] [blame] | 1066 | #endif |
| 1067 | |
Bernhard Reutner-Fischer | 7e8a53a | 2007-04-10 09:37:29 +0000 | [diff] [blame] | 1068 | /* Determine where to start transfer */ |
Denys Vlasenko | 2384a35 | 2011-02-15 00:58:36 +0100 | [diff] [blame] | 1069 | G.beg_range = 0; |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1070 | if (option_mask32 & WGET_OPT_CONTINUE) { |
Denys Vlasenko | 2384a35 | 2011-02-15 00:58:36 +0100 | [diff] [blame] | 1071 | G.output_fd = open(G.fname_out, O_WRONLY); |
| 1072 | if (G.output_fd >= 0) { |
| 1073 | G.beg_range = xlseek(G.output_fd, 0, SEEK_END); |
Denis Vlasenko | a94554d | 2006-09-23 17:49:09 +0000 | [diff] [blame] | 1074 | } |
| 1075 | /* File doesn't exist. We do not create file here yet. |
Denys Vlasenko | a84eadf | 2011-02-12 23:40:31 +0100 | [diff] [blame] | 1076 | * We are not sure it exists on remote side */ |
Eric Andersen | 9670083 | 2000-09-04 15:15:55 +0000 | [diff] [blame] | 1077 | } |
| 1078 | |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1079 | redir_limit = 5; |
| 1080 | resolve_lsa: |
Denis Vlasenko | 42823d5 | 2007-02-04 02:39:08 +0000 | [diff] [blame] | 1081 | lsa = xhost2sockaddr(server.host, server.port); |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1082 | if (!(option_mask32 & WGET_OPT_QUIET)) { |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1083 | char *s = xmalloc_sockaddr2dotted(&lsa->u.sa); |
| 1084 | fprintf(stderr, "Connecting to %s (%s)\n", server.host, s); |
| 1085 | free(s); |
Eric Andersen | e6dc439 | 2003-10-31 09:31:46 +0000 | [diff] [blame] | 1086 | } |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1087 | establish_session: |
Denys Vlasenko | 2384a35 | 2011-02-15 00:58:36 +0100 | [diff] [blame] | 1088 | /*G.content_len = 0; - redundant, got_clen = 0 is enough */ |
| 1089 | G.got_clen = 0; |
| 1090 | G.chunked = 0; |
Denys Vlasenko | 403f299 | 2018-02-06 15:15:08 +0100 | [diff] [blame] | 1091 | if (use_proxy || target.protocol[0] != 'f' /*not ftp[s]*/) { |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 1092 | /* |
| 1093 | * HTTP session |
| 1094 | */ |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1095 | char *str; |
Denys Vlasenko | 7f43280 | 2009-06-28 01:02:24 +0200 | [diff] [blame] | 1096 | int status; |
Denys Vlasenko | 7f43280 | 2009-06-28 01:02:24 +0200 | [diff] [blame] | 1097 | |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 1098 | /* Open socket to http(s) server */ |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 1099 | #if ENABLE_FEATURE_WGET_OPENSSL |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 1100 | /* openssl (and maybe internal TLS) support is configured */ |
Peter Lloyd | 804ce5a | 2018-03-05 00:17:02 +0100 | [diff] [blame] | 1101 | if (server.protocol == P_HTTPS) { |
Denys Vlasenko | 1c6c670 | 2015-10-07 01:39:40 +0200 | [diff] [blame] | 1102 | /* openssl-based helper |
| 1103 | * Inconvenient API since we can't give it an open fd |
| 1104 | */ |
Denys Vlasenko | 2007ef5 | 2015-10-07 02:40:53 +0200 | [diff] [blame] | 1105 | int fd = spawn_https_helper_openssl(server.host, server.port); |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 1106 | # if ENABLE_FEATURE_WGET_HTTPS |
| 1107 | if (fd < 0) { /* no openssl? try internal */ |
Denys Vlasenko | 2007ef5 | 2015-10-07 02:40:53 +0200 | [diff] [blame] | 1108 | sfp = open_socket(lsa); |
Denys Vlasenko | 403f299 | 2018-02-06 15:15:08 +0100 | [diff] [blame] | 1109 | spawn_ssl_client(server.host, fileno(sfp), /*flags*/ 0); |
Denys Vlasenko | 2007ef5 | 2015-10-07 02:40:53 +0200 | [diff] [blame] | 1110 | goto socket_opened; |
| 1111 | } |
| 1112 | # else |
| 1113 | /* We don't check for exec("openssl") failure in this case */ |
| 1114 | # endif |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 1115 | sfp = fdopen(fd, "r+"); |
| 1116 | if (!sfp) |
| 1117 | bb_perror_msg_and_die(bb_msg_memory_exhausted); |
Denys Vlasenko | 2007ef5 | 2015-10-07 02:40:53 +0200 | [diff] [blame] | 1118 | goto socket_opened; |
| 1119 | } |
| 1120 | sfp = open_socket(lsa); |
| 1121 | socket_opened: |
Denys Vlasenko | 9a647c3 | 2017-01-23 01:08:16 +0100 | [diff] [blame] | 1122 | #elif ENABLE_FEATURE_WGET_HTTPS |
| 1123 | /* Only internal TLS support is configured */ |
Denys Vlasenko | 2007ef5 | 2015-10-07 02:40:53 +0200 | [diff] [blame] | 1124 | sfp = open_socket(lsa); |
Peter Lloyd | 804ce5a | 2018-03-05 00:17:02 +0100 | [diff] [blame] | 1125 | if (server.protocol == P_HTTPS) |
Denys Vlasenko | 403f299 | 2018-02-06 15:15:08 +0100 | [diff] [blame] | 1126 | spawn_ssl_client(server.host, fileno(sfp), /*flags*/ 0); |
Denys Vlasenko | 2007ef5 | 2015-10-07 02:40:53 +0200 | [diff] [blame] | 1127 | #else |
| 1128 | /* ssl (https) support is not configured */ |
| 1129 | sfp = open_socket(lsa); |
Denys Vlasenko | 5331557 | 2014-02-23 23:39:47 +0100 | [diff] [blame] | 1130 | #endif |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1131 | /* Send HTTP request */ |
| 1132 | if (use_proxy) { |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 1133 | SENDFMT(sfp, "GET %s://%s/%s HTTP/1.1\r\n", |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 1134 | target.protocol, target.host, |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1135 | target.path); |
| 1136 | } else { |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 1137 | SENDFMT(sfp, "%s /%s HTTP/1.1\r\n", |
Denys Vlasenko | 8b7e8ae | 2014-02-22 14:12:29 +0100 | [diff] [blame] | 1138 | (option_mask32 & WGET_OPT_POST_DATA) ? "POST" : "GET", |
| 1139 | target.path); |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1140 | } |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 1141 | if (!USR_HEADER_HOST) |
| 1142 | SENDFMT(sfp, "Host: %s\r\n", target.host); |
| 1143 | if (!USR_HEADER_USER_AGENT) |
| 1144 | SENDFMT(sfp, "User-Agent: %s\r\n", G.user_agent); |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 1145 | |
Denys Vlasenko | 9213a55 | 2011-02-10 13:23:45 +0100 | [diff] [blame] | 1146 | /* Ask server to close the connection as soon as we are done |
| 1147 | * (IOW: we do not intend to send more requests) |
| 1148 | */ |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 1149 | SENDFMT(sfp, "Connection: close\r\n"); |
Denys Vlasenko | 9213a55 | 2011-02-10 13:23:45 +0100 | [diff] [blame] | 1150 | |
Denis Vlasenko | 9cade08 | 2006-11-21 10:43:02 +0000 | [diff] [blame] | 1151 | #if ENABLE_FEATURE_WGET_AUTHENTICATION |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 1152 | if (target.user && !USR_HEADER_AUTH) { |
| 1153 | SENDFMT(sfp, "Proxy-Authorization: Basic %s\r\n"+6, |
Denys Vlasenko | df4e16c | 2011-02-10 06:29:06 +0100 | [diff] [blame] | 1154 | base64enc(target.user)); |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1155 | } |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 1156 | if (use_proxy && server.user && !USR_HEADER_PROXY_AUTH) { |
| 1157 | SENDFMT(sfp, "Proxy-Authorization: Basic %s\r\n", |
Denys Vlasenko | df4e16c | 2011-02-10 06:29:06 +0100 | [diff] [blame] | 1158 | base64enc(server.user)); |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1159 | } |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 1160 | #endif |
| 1161 | |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 1162 | if (G.beg_range != 0 && !USR_HEADER_RANGE) |
| 1163 | SENDFMT(sfp, "Range: bytes=%"OFF_FMT"u-\r\n", G.beg_range); |
Denys Vlasenko | 9213a55 | 2011-02-10 13:23:45 +0100 | [diff] [blame] | 1164 | |
Denis Vlasenko | c8400a2 | 2006-10-25 00:33:44 +0000 | [diff] [blame] | 1165 | #if ENABLE_FEATURE_WGET_LONG_OPTIONS |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 1166 | if (G.extra_headers) { |
| 1167 | log_io(G.extra_headers); |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1168 | fputs(G.extra_headers, sfp); |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 1169 | } |
Denis Vlasenko | 5a2ad69 | 2009-03-04 14:13:37 +0000 | [diff] [blame] | 1170 | |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1171 | if (option_mask32 & WGET_OPT_POST_DATA) { |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 1172 | SENDFMT(sfp, |
Denys Vlasenko | 9213a55 | 2011-02-10 13:23:45 +0100 | [diff] [blame] | 1173 | "Content-Type: application/x-www-form-urlencoded\r\n" |
| 1174 | "Content-Length: %u\r\n" |
| 1175 | "\r\n" |
| 1176 | "%s", |
Vitaly Magerya | 700fbc3 | 2011-03-27 22:33:13 +0200 | [diff] [blame] | 1177 | (int) strlen(G.post_data), G.post_data |
Denys Vlasenko | 9213a55 | 2011-02-10 13:23:45 +0100 | [diff] [blame] | 1178 | ); |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1179 | } else |
Denis Vlasenko | c8400a2 | 2006-10-25 00:33:44 +0000 | [diff] [blame] | 1180 | #endif |
Denys Vlasenko | 9213a55 | 2011-02-10 13:23:45 +0100 | [diff] [blame] | 1181 | { |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 1182 | SENDFMT(sfp, "\r\n"); |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1183 | } |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 1184 | |
Nguyễn Thái Ngọc Duy | ebec11d | 2010-09-23 15:18:41 +0200 | [diff] [blame] | 1185 | fflush(sfp); |
Denys Vlasenko | a6f8651 | 2017-01-11 20:16:45 +0100 | [diff] [blame] | 1186 | |
Denys Vlasenko | 4e08a12 | 2017-01-16 17:31:05 +0100 | [diff] [blame] | 1187 | /* Tried doing this unconditionally. |
| 1188 | * Cloudflare and nginx/1.11.5 are shocked to see SHUT_WR on non-HTTPS. |
| 1189 | */ |
Denys Vlasenko | a6f8651 | 2017-01-11 20:16:45 +0100 | [diff] [blame] | 1190 | #if SSL_SUPPORTED |
| 1191 | if (target.protocol == P_HTTPS) { |
| 1192 | /* If we use SSL helper, keeping our end of the socket open for writing |
| 1193 | * makes our end (i.e. the same fd!) readable (EAGAIN instead of EOF) |
| 1194 | * even after child closes its copy of the fd. |
| 1195 | * This helps: |
| 1196 | */ |
| 1197 | shutdown(fileno(sfp), SHUT_WR); |
| 1198 | } |
| 1199 | #endif |
Nguyễn Thái Ngọc Duy | ebec11d | 2010-09-23 15:18:41 +0200 | [diff] [blame] | 1200 | |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1201 | /* |
| 1202 | * Retrieve HTTP response line and check for "200" status code. |
| 1203 | */ |
Denis Vlasenko | 023b57d | 2006-10-15 17:05:55 +0000 | [diff] [blame] | 1204 | read_response: |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 1205 | fgets_trim_sanitize(sfp, " %s\n"); |
Eric Andersen | c7bda1c | 2004-03-15 08:29:22 +0000 | [diff] [blame] | 1206 | |
Denys Vlasenko | df4e16c | 2011-02-10 06:29:06 +0100 | [diff] [blame] | 1207 | str = G.wget_buf; |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1208 | str = skip_non_whitespace(str); |
| 1209 | str = skip_whitespace(str); |
| 1210 | // FIXME: no error check |
| 1211 | // xatou wouldn't work: "200 OK" |
| 1212 | status = atoi(str); |
| 1213 | switch (status) { |
| 1214 | case 0: |
| 1215 | case 100: |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 1216 | while (get_sanitized_hdr(sfp) != NULL) |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1217 | /* eat all remaining headers */; |
| 1218 | goto read_response; |
Denys Vlasenko | 9ff910d | 2016-08-31 13:28:53 +0200 | [diff] [blame] | 1219 | |
| 1220 | /* Success responses */ |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1221 | case 200: |
Denys Vlasenko | 9ff910d | 2016-08-31 13:28:53 +0200 | [diff] [blame] | 1222 | /* fall through */ |
| 1223 | case 201: /* 201 Created */ |
| 1224 | /* "The request has been fulfilled and resulted in a new resource being created" */ |
Denys Vlasenko | ef15970 | 2016-09-01 11:16:22 +0200 | [diff] [blame] | 1225 | /* Standard wget is reported to treat this as success */ |
Denys Vlasenko | 9ff910d | 2016-08-31 13:28:53 +0200 | [diff] [blame] | 1226 | /* fall through */ |
| 1227 | case 202: /* 202 Accepted */ |
| 1228 | /* "The request has been accepted for processing, but the processing has not been completed" */ |
| 1229 | /* Treat as success: fall through */ |
| 1230 | case 203: /* 203 Non-Authoritative Information */ |
| 1231 | /* "Use of this response code is not required and is only appropriate when the response would otherwise be 200 (OK)" */ |
| 1232 | /* fall through */ |
| 1233 | case 204: /* 204 No Content */ |
Denis Vlasenko | 50b5cac | 2008-06-22 16:28:02 +0000 | [diff] [blame] | 1234 | /* |
| 1235 | Response 204 doesn't say "null file", it says "metadata |
| 1236 | has changed but data didn't": |
| 1237 | |
| 1238 | "10.2.5 204 No Content |
| 1239 | The server has fulfilled the request but does not need to return |
| 1240 | an entity-body, and might want to return updated metainformation. |
| 1241 | The response MAY include new or updated metainformation in the form |
| 1242 | of entity-headers, which if present SHOULD be associated with |
| 1243 | the requested variant. |
| 1244 | |
| 1245 | If the client is a user agent, it SHOULD NOT change its document |
| 1246 | view from that which caused the request to be sent. This response |
| 1247 | is primarily intended to allow input for actions to take place |
| 1248 | without causing a change to the user agent's active document view, |
| 1249 | although any new or updated metainformation SHOULD be applied |
| 1250 | to the document currently in the user agent's active view. |
| 1251 | |
| 1252 | The 204 response MUST NOT include a message-body, and thus |
| 1253 | is always terminated by the first empty line after the header fields." |
| 1254 | |
| 1255 | However, in real world it was observed that some web servers |
| 1256 | (e.g. Boa/0.94.14rc21) simply use code 204 when file size is zero. |
| 1257 | */ |
Denys Vlasenko | bf146b8 | 2012-06-13 17:31:07 +0200 | [diff] [blame] | 1258 | if (G.beg_range != 0) { |
| 1259 | /* "Range:..." was not honored by the server. |
| 1260 | * Restart download from the beginning. |
| 1261 | */ |
| 1262 | reset_beg_range_to_zero(); |
| 1263 | } |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1264 | break; |
Denys Vlasenko | 9ff910d | 2016-08-31 13:28:53 +0200 | [diff] [blame] | 1265 | /* 205 Reset Content ?? what to do on this ?? */ |
| 1266 | |
Denys Vlasenko | fb132e4 | 2010-10-29 11:46:52 +0200 | [diff] [blame] | 1267 | case 300: /* redirection */ |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1268 | case 301: |
| 1269 | case 302: |
| 1270 | case 303: |
| 1271 | break; |
Denys Vlasenko | 9ff910d | 2016-08-31 13:28:53 +0200 | [diff] [blame] | 1272 | |
Vladimir Dronnikov | f5abc78 | 2012-06-13 17:29:41 +0200 | [diff] [blame] | 1273 | case 206: /* Partial Content */ |
| 1274 | if (G.beg_range != 0) |
| 1275 | /* "Range:..." worked. Good. */ |
Denis Vlasenko | 023b57d | 2006-10-15 17:05:55 +0000 | [diff] [blame] | 1276 | break; |
Vladimir Dronnikov | f5abc78 | 2012-06-13 17:29:41 +0200 | [diff] [blame] | 1277 | /* Partial Content even though we did not ask for it??? */ |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1278 | /* fall through */ |
| 1279 | default: |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 1280 | bb_error_msg_and_die("server returned error: %s", G.wget_buf); |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1281 | } |
Eric Andersen | c7bda1c | 2004-03-15 08:29:22 +0000 | [diff] [blame] | 1282 | |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1283 | /* |
| 1284 | * Retrieve HTTP headers. |
| 1285 | */ |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 1286 | while ((str = get_sanitized_hdr(sfp)) != NULL) { |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1287 | static const char keywords[] ALIGN1 = |
| 1288 | "content-length\0""transfer-encoding\0""location\0"; |
| 1289 | enum { |
| 1290 | KEY_content_length = 1, KEY_transfer_encoding, KEY_location |
| 1291 | }; |
Matthijs van de Water | 0d58666 | 2009-08-22 20:19:48 +0200 | [diff] [blame] | 1292 | smalluint key; |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1293 | |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 1294 | /* get_sanitized_hdr converted "FOO:" string to lowercase */ |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1295 | |
Matthijs van de Water | 0d58666 | 2009-08-22 20:19:48 +0200 | [diff] [blame] | 1296 | /* strip trailing whitespace */ |
| 1297 | char *s = strchrnul(str, '\0') - 1; |
| 1298 | while (s >= str && (*s == ' ' || *s == '\t')) { |
| 1299 | *s = '\0'; |
| 1300 | s--; |
| 1301 | } |
Denys Vlasenko | df4e16c | 2011-02-10 06:29:06 +0100 | [diff] [blame] | 1302 | key = index_in_strings(keywords, G.wget_buf) + 1; |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1303 | if (key == KEY_content_length) { |
Denys Vlasenko | a3aa3e3 | 2009-12-11 12:36:10 +0100 | [diff] [blame] | 1304 | G.content_len = BB_STRTOOFF(str, NULL, 10); |
| 1305 | if (G.content_len < 0 || errno) { |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 1306 | bb_error_msg_and_die("content-length %s is garbage", str); |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 1307 | } |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1308 | G.got_clen = 1; |
| 1309 | continue; |
| 1310 | } |
| 1311 | if (key == KEY_transfer_encoding) { |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1312 | if (strcmp(str_tolower(str), "chunked") != 0) |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 1313 | bb_error_msg_and_die("transfer encoding '%s' is not supported", str); |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1314 | G.chunked = 1; |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1315 | } |
| 1316 | if (key == KEY_location && status >= 300) { |
| 1317 | if (--redir_limit == 0) |
| 1318 | bb_error_msg_and_die("too many redirections"); |
| 1319 | fclose(sfp); |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1320 | if (str[0] == '/') { |
Denys Vlasenko | 93b4a60 | 2011-12-18 05:11:56 +0100 | [diff] [blame] | 1321 | free(redirected_path); |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 1322 | target.path = redirected_path = xstrdup(str + 1); |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1323 | /* lsa stays the same: it's on the same server */ |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1324 | } else { |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1325 | parse_url(str, &target); |
| 1326 | if (!use_proxy) { |
Denys Vlasenko | d353bff | 2014-02-03 14:09:42 +0100 | [diff] [blame] | 1327 | /* server.user remains untouched */ |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1328 | free(server.allocated); |
Pere Orga | 57b4909 | 2011-02-14 23:56:07 +0100 | [diff] [blame] | 1329 | server.allocated = NULL; |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1330 | server.host = target.host; |
Denys Vlasenko | 7d5ddf1 | 2009-06-30 20:36:27 +0200 | [diff] [blame] | 1331 | /* strip_ipv6_scope_id(target.host); - no! */ |
| 1332 | /* we assume remote never gives us IPv6 addr with scope id */ |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1333 | server.port = target.port; |
Denis Vlasenko | 6536a9b | 2007-01-12 10:35:23 +0000 | [diff] [blame] | 1334 | free(lsa); |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1335 | goto resolve_lsa; |
| 1336 | } /* else: lsa stays the same: we use proxy */ |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 1337 | } |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1338 | goto establish_session; |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 1339 | } |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1340 | } |
| 1341 | // if (status >= 300) |
| 1342 | // bb_error_msg_and_die("bad redirection (no Location: header from server)"); |
Eric Andersen | c7bda1c | 2004-03-15 08:29:22 +0000 | [diff] [blame] | 1343 | |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1344 | /* For HTTP, data is pumped over the same connection */ |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 1345 | dfp = sfp; |
Denis Vlasenko | 96e9d3c | 2006-10-07 14:28:55 +0000 | [diff] [blame] | 1346 | } else { |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 1347 | /* |
| 1348 | * FTP session |
| 1349 | */ |
Denys Vlasenko | 7f43280 | 2009-06-28 01:02:24 +0200 | [diff] [blame] | 1350 | sfp = prepare_ftp_session(&dfp, &target, lsa); |
Eric Andersen | 9670083 | 2000-09-04 15:15:55 +0000 | [diff] [blame] | 1351 | } |
Denis Vlasenko | 7710563 | 2007-09-24 15:04:00 +0000 | [diff] [blame] | 1352 | |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1353 | free(lsa); |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1354 | |
Denys Vlasenko | 9a5b7f6 | 2011-02-13 02:49:43 +0100 | [diff] [blame] | 1355 | if (!(option_mask32 & WGET_OPT_SPIDER)) { |
Denys Vlasenko | 2384a35 | 2011-02-15 00:58:36 +0100 | [diff] [blame] | 1356 | if (G.output_fd < 0) |
| 1357 | G.output_fd = xopen(G.fname_out, G.o_flags); |
| 1358 | retrieve_file_data(dfp); |
| 1359 | if (!(option_mask32 & WGET_OPT_OUTNAME)) { |
| 1360 | xclose(G.output_fd); |
| 1361 | G.output_fd = -1; |
Denys Vlasenko | 9a5b7f6 | 2011-02-13 02:49:43 +0100 | [diff] [blame] | 1362 | } |
Bernhard Reutner-Fischer | 2e75dcc | 2007-04-05 10:31:47 +0000 | [diff] [blame] | 1363 | } |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 1364 | |
Denys Vlasenko | f1fab09 | 2009-06-28 03:33:57 +0200 | [diff] [blame] | 1365 | if (dfp != sfp) { |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1366 | /* It's ftp. Close data connection properly */ |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 1367 | fclose(dfp); |
Denys Vlasenko | df4e16c | 2011-02-10 06:29:06 +0100 | [diff] [blame] | 1368 | if (ftpcmd(NULL, NULL, sfp) != 226) |
Denys Vlasenko | 3459024 | 2018-02-12 16:46:13 +0100 | [diff] [blame] | 1369 | bb_error_msg_and_die("ftp error: %s", G.wget_buf); |
Denys Vlasenko | df4e16c | 2011-02-10 06:29:06 +0100 | [diff] [blame] | 1370 | /* ftpcmd("QUIT", NULL, sfp); - why bother? */ |
Eric Andersen | 79757c9 | 2001-04-05 21:45:54 +0000 | [diff] [blame] | 1371 | } |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1372 | fclose(sfp); |
Denis Vlasenko | 7710563 | 2007-09-24 15:04:00 +0000 | [diff] [blame] | 1373 | |
Denys Vlasenko | 9a5b7f6 | 2011-02-13 02:49:43 +0100 | [diff] [blame] | 1374 | free(server.allocated); |
| 1375 | free(target.allocated); |
Denys Vlasenko | d353bff | 2014-02-03 14:09:42 +0100 | [diff] [blame] | 1376 | free(server.user); |
| 1377 | free(target.user); |
Denys Vlasenko | 9a5b7f6 | 2011-02-13 02:49:43 +0100 | [diff] [blame] | 1378 | free(fname_out_alloc); |
Denys Vlasenko | 93b4a60 | 2011-12-18 05:11:56 +0100 | [diff] [blame] | 1379 | free(redirected_path); |
Eric Andersen | 9670083 | 2000-09-04 15:15:55 +0000 | [diff] [blame] | 1380 | } |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1381 | |
| 1382 | int wget_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE; |
| 1383 | int wget_main(int argc UNUSED_PARAM, char **argv) |
| 1384 | { |
| 1385 | #if ENABLE_FEATURE_WGET_LONG_OPTIONS |
| 1386 | static const char wget_longopts[] ALIGN1 = |
| 1387 | /* name, has_arg, val */ |
| 1388 | "continue\0" No_argument "c" |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1389 | "quiet\0" No_argument "q" |
Denys Vlasenko | dff9fef | 2017-01-24 21:41:43 +0100 | [diff] [blame] | 1390 | "server-response\0" No_argument "S" |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1391 | "output-document\0" Required_argument "O" |
| 1392 | "directory-prefix\0" Required_argument "P" |
| 1393 | "proxy\0" Required_argument "Y" |
| 1394 | "user-agent\0" Required_argument "U" |
Denys Vlasenko | 92e1b08 | 2015-10-20 21:51:52 +0200 | [diff] [blame] | 1395 | IF_FEATURE_WGET_TIMEOUT( |
| 1396 | "timeout\0" Required_argument "T") |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1397 | /* Ignored: */ |
Denys Vlasenko | 92e1b08 | 2015-10-20 21:51:52 +0200 | [diff] [blame] | 1398 | IF_DESKTOP( "tries\0" Required_argument "t") |
| 1399 | "header\0" Required_argument "\xff" |
| 1400 | "post-data\0" Required_argument "\xfe" |
Denys Vlasenko | 2972e2c | 2016-10-04 04:23:09 +0200 | [diff] [blame] | 1401 | "spider\0" No_argument "\xfd" |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1402 | /* Ignored (we always use PASV): */ |
Denys Vlasenko | 92e1b08 | 2015-10-20 21:51:52 +0200 | [diff] [blame] | 1403 | IF_DESKTOP( "passive-ftp\0" No_argument "\xf0") |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1404 | /* Ignored (we don't do ssl) */ |
Denys Vlasenko | 92e1b08 | 2015-10-20 21:51:52 +0200 | [diff] [blame] | 1405 | IF_DESKTOP( "no-check-certificate\0" No_argument "\xf0") |
Vladimir Dronnikov | f5abc78 | 2012-06-13 17:29:41 +0200 | [diff] [blame] | 1406 | /* Ignored (we don't support caching) */ |
Denys Vlasenko | 92e1b08 | 2015-10-20 21:51:52 +0200 | [diff] [blame] | 1407 | IF_DESKTOP( "no-cache\0" No_argument "\xf0") |
| 1408 | IF_DESKTOP( "no-verbose\0" No_argument "\xf0") |
| 1409 | IF_DESKTOP( "no-clobber\0" No_argument "\xf0") |
| 1410 | IF_DESKTOP( "no-host-directories\0" No_argument "\xf0") |
| 1411 | IF_DESKTOP( "no-parent\0" No_argument "\xf0") |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1412 | ; |
Denys Vlasenko | 036585a | 2017-08-08 16:38:18 +0200 | [diff] [blame] | 1413 | # define GETOPT32 getopt32long |
| 1414 | # define LONGOPTS ,wget_longopts |
| 1415 | #else |
| 1416 | # define GETOPT32 getopt32 |
| 1417 | # define LONGOPTS |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1418 | #endif |
| 1419 | |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1420 | #if ENABLE_FEATURE_WGET_LONG_OPTIONS |
| 1421 | llist_t *headers_llist = NULL; |
| 1422 | #endif |
| 1423 | |
| 1424 | INIT_G(); |
| 1425 | |
Lauri Kasanen | d074b41 | 2013-10-12 21:47:07 +0200 | [diff] [blame] | 1426 | #if ENABLE_FEATURE_WGET_TIMEOUT |
| 1427 | G.timeout_seconds = 900; |
| 1428 | signal(SIGALRM, alarm_handler); |
| 1429 | #endif |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1430 | G.proxy_flag = "on"; /* use proxies if env vars are set */ |
| 1431 | G.user_agent = "Wget"; /* "User-Agent" header field */ |
| 1432 | |
| 1433 | #if ENABLE_FEATURE_WGET_LONG_OPTIONS |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1434 | #endif |
Denys Vlasenko | 22542ec | 2017-08-08 21:55:02 +0200 | [diff] [blame] | 1435 | GETOPT32(argv, "^" |
| 1436 | "cqSO:P:Y:U:T:+" |
Denys Vlasenko | 92e1b08 | 2015-10-20 21:51:52 +0200 | [diff] [blame] | 1437 | /*ignored:*/ "t:" |
| 1438 | /*ignored:*/ "n::" |
| 1439 | /* wget has exactly four -n<letter> opts, all of which we can ignore: |
| 1440 | * -nv --no-verbose: be moderately quiet (-q is full quiet) |
| 1441 | * -nc --no-clobber: abort if exists, neither download to FILE.n nor overwrite FILE |
| 1442 | * -nH --no-host-directories: wget -r http://host/ won't create host/ |
| 1443 | * -np --no-parent |
| 1444 | * "n::" above says that we accept -n[ARG]. |
| 1445 | * Specifying "n:" would be a bug: "-n ARG" would eat ARG! |
| 1446 | */ |
Denys Vlasenko | 22542ec | 2017-08-08 21:55:02 +0200 | [diff] [blame] | 1447 | "\0" |
| 1448 | "-1" /* at least one URL */ |
| 1449 | IF_FEATURE_WGET_LONG_OPTIONS(":\xff::") /* --header is a list */ |
Denys Vlasenko | 036585a | 2017-08-08 16:38:18 +0200 | [diff] [blame] | 1450 | LONGOPTS |
Denys Vlasenko | 92e1b08 | 2015-10-20 21:51:52 +0200 | [diff] [blame] | 1451 | , &G.fname_out, &G.dir_prefix, |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1452 | &G.proxy_flag, &G.user_agent, |
| 1453 | IF_FEATURE_WGET_TIMEOUT(&G.timeout_seconds) IF_NOT_FEATURE_WGET_TIMEOUT(NULL), |
Denys Vlasenko | 92e1b08 | 2015-10-20 21:51:52 +0200 | [diff] [blame] | 1454 | NULL, /* -t RETRIES */ |
| 1455 | NULL /* -n[ARG] */ |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1456 | IF_FEATURE_WGET_LONG_OPTIONS(, &headers_llist) |
| 1457 | IF_FEATURE_WGET_LONG_OPTIONS(, &G.post_data) |
| 1458 | ); |
Denys Vlasenko | 2972e2c | 2016-10-04 04:23:09 +0200 | [diff] [blame] | 1459 | #if 0 /* option bits debug */ |
| 1460 | if (option_mask32 & WGET_OPT_RETRIES) bb_error_msg("-t NUM"); |
| 1461 | if (option_mask32 & WGET_OPT_nsomething) bb_error_msg("-nsomething"); |
| 1462 | if (option_mask32 & WGET_OPT_HEADER) bb_error_msg("--header"); |
| 1463 | if (option_mask32 & WGET_OPT_POST_DATA) bb_error_msg("--post-data"); |
| 1464 | if (option_mask32 & WGET_OPT_SPIDER) bb_error_msg("--spider"); |
| 1465 | exit(0); |
| 1466 | #endif |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1467 | argv += optind; |
| 1468 | |
| 1469 | #if ENABLE_FEATURE_WGET_LONG_OPTIONS |
| 1470 | if (headers_llist) { |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 1471 | int size = 0; |
| 1472 | char *hdr; |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1473 | llist_t *ll = headers_llist; |
| 1474 | while (ll) { |
| 1475 | size += strlen(ll->data) + 2; |
| 1476 | ll = ll->link; |
| 1477 | } |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 1478 | G.extra_headers = hdr = xmalloc(size + 1); |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1479 | while (headers_llist) { |
Bernhard Reutner-Fischer | d7bfee1 | 2015-02-18 20:41:02 +0100 | [diff] [blame] | 1480 | int bit; |
| 1481 | const char *words; |
| 1482 | |
| 1483 | size = sprintf(hdr, "%s\r\n", |
| 1484 | (char*)llist_pop(&headers_llist)); |
| 1485 | /* a bit like index_in_substrings but don't match full key */ |
| 1486 | bit = 1; |
| 1487 | words = wget_user_headers; |
| 1488 | while (*words) { |
| 1489 | if (strstr(hdr, words) == hdr) { |
| 1490 | G.user_headers |= bit; |
| 1491 | break; |
| 1492 | } |
| 1493 | bit <<= 1; |
| 1494 | words += strlen(words) + 1; |
| 1495 | } |
| 1496 | hdr += size; |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1497 | } |
| 1498 | } |
| 1499 | #endif |
| 1500 | |
Denys Vlasenko | 2384a35 | 2011-02-15 00:58:36 +0100 | [diff] [blame] | 1501 | G.output_fd = -1; |
| 1502 | G.o_flags = O_WRONLY | O_CREAT | O_TRUNC | O_EXCL; |
| 1503 | if (G.fname_out) { /* -O FILE ? */ |
| 1504 | if (LONE_DASH(G.fname_out)) { /* -O - ? */ |
| 1505 | G.output_fd = 1; |
| 1506 | option_mask32 &= ~WGET_OPT_CONTINUE; |
| 1507 | } |
| 1508 | /* compat with wget: -O FILE can overwrite */ |
| 1509 | G.o_flags = O_WRONLY | O_CREAT | O_TRUNC; |
| 1510 | } |
| 1511 | |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1512 | while (*argv) |
Pere Orga | 5369563 | 2011-02-16 20:09:36 +0100 | [diff] [blame] | 1513 | download_one_url(*argv++); |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1514 | |
Denys Vlasenko | 28556b9 | 2011-02-15 11:03:53 +0100 | [diff] [blame] | 1515 | if (G.output_fd >= 0) |
| 1516 | xclose(G.output_fd); |
| 1517 | |
Guilherme Maciel Ferreira | 840ef17 | 2013-10-16 14:43:30 +0200 | [diff] [blame] | 1518 | #if ENABLE_FEATURE_CLEAN_UP && ENABLE_FEATURE_WGET_LONG_OPTIONS |
| 1519 | free(G.extra_headers); |
| 1520 | #endif |
| 1521 | FINI_G(); |
| 1522 | |
Pere Orga | 5369563 | 2011-02-16 20:09:36 +0100 | [diff] [blame] | 1523 | return EXIT_SUCCESS; |
Denys Vlasenko | a366109 | 2011-02-13 02:33:11 +0100 | [diff] [blame] | 1524 | } |