1 GNU Wget NEWS -- history of user-visible changes.
3 Copyright (C) 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005,
4 2006, 2007, 2008 Free Software Foundation, Inc.
5 See the end for copying conditions.
7 Please send GNU Wget bug reports to <bug-wget@gnu.org>.
9 * Changes in Wget (MAINLINE).
11 ** Gnulib is now used to provide code for certain portability aspects in
14 ** Wget uses Automake now as part of its build infrastructure.
16 * Changes in Wget 1.11.2
18 ** Fixed a problem in authenticating over HTTPS through a proxy.
19 (Regression in 1.11 over 1.10.2.)
21 ** The combination of -r or -p with -O, which was disallowed in 1.11,
22 has been downgraded to a warning in 1.11.2. (-O and -N, which was never
23 meaningful, is still an error.)
25 ** Further improvements to progress bar displays in non-English locales
26 (too many spaces could be inserted, causing the display to scroll).
28 ** Successive invocations of Wget on FTP URLS, with --no-remove-listing
29 and --continue, was causing Wget to append, rather than replace,
30 information in the .listing file, and thereby download the same files
31 multiple times. This has been fixed in 1.11.2.
33 ** Wget 1.11 no longer allowed ".." to persist at the beginning of URLs,
34 for improved conformance with RFC 3986. However, this behavior presents
35 problems for some FTP setups, and so they are now preserved again, for
38 * Changes in Wget 1.11.1.
40 ** Interrupted downloads no longer result in renaming the file
41 (regression in 1.11 over 1.10.2).
43 ** Progress bar now displays correctly in non-English locales (and a
44 related assertion failure was fixed).
46 ** Wget no longer issues a GET request over HTTP for files it should
47 know it's not going to download (regression in 1.11 over 1.10.2).
49 ** Added option --auth-no-challenge, to support broken pre-1.11
50 authentication-before-server-challenge, which turns out to still be
51 useful for some limited cases.
53 ** Documentation of accept/reject lists in the manual's "Types of
54 Files" section now explains various aspects of their behavior that may
55 be surprising, and notes that they may change in the future.
57 ** Documentation of --no-parents now explains how a trailing slash, or
58 lack thereof, in the specified URL, will affect behavior.
60 * Changes in Wget 1.11.
62 ** Timestamping now uses the value from the most recent HTTP response,
63 rather than the first one it got.
65 ** Authentication information is no longer sent as part of the Referer
66 header in recursive fetches.
68 ** No authentication credentials are sent until a challenge is issued,
69 for improved security. Authentication handling is still not
70 RFC-compliant, as once a Basic challenge has been received, it will
71 assume it can send credentials to any URL at that same host, and not
72 just the ones at or below the original authenticated location.
73 Credentials for Digest authentication are still never saved or issued
74 automatically, and continue to require a challenge for each resource.
76 ** Added --max-redirect option, allowing the user to specify what should
77 be the maximum number of HTTP redirects to follow.
79 ** Wget now supports saving HTTP downloads using file names specified by
80 the `Content-Disposition' header. This is a standard way of specifying
81 the file name used by many web dynamically generated pages. However, the
82 current implementation is inefficient, and known to have bugs. It is
83 EXPERIMENTAL only, and not enabled by default. Use --content-disposition
86 ** The new option `--ignore-case' makes Wget ignore case when
87 matching files, directories, and wildcards. This affects the -X, -I,
88 -A, and -R options, as well as globbing in FTP URLs.
90 ** ETA projection is now displayed in "dot" progress output as well as
91 in the default progress bar. (The dot progress is used by default when
92 logging Wget's output to file using the `-o' option.)
94 ** The "lockable boolean" argument type is no longer supported. It
95 was only used by the passive_ftp .wgetrc setting. If you're running
96 broken scripts or Perl modules that unconditionally specify
97 `--passive-ftp' and your firewall disallows it, you can override them
98 by replacing wget with a script that execs wget "$@" --no-passive-ftp.
100 ** The source code has been migrated to Mercurial. The repositories are
101 available at http://hg.addictivecode.org/. Prior to this, the source
102 code was hosted on Subversion (migrated from the original CVS); you can
103 still get access to older tags and branches for Wget in the Subversion
104 repository at http://addictivecode.org/svn/wget/.
106 * Changes in Wget 1.10.
108 ** Downloading files larger than 2GB, sometimes referred to as "large
109 files", now works on systems that support them. This includes the
110 majority of modern Unixes, as well as MS Windows.
112 ** IPv6 is now supported by Wget. Unlike the experimental code in
113 1.9, this version supports dual-family systems. The new flags
114 `--inet4' and `--inet6' (or `-4' and `-6' for short) force the use of
115 IPv4 and IPv6 respectively. Note that IPv6 support has not yet been
118 ** Microsoft's proprietary "NTLM" method of HTTP authentication is now
119 supported. This authentication method is undocumented and only used
120 by IIS. Note that *proxy* authentication is not supported in this
121 release; you can only authenticate to the target web site.
123 ** Wget no longer truncates partially downloaded files when download
124 has to start over because the server doesn't support Range. Instead,
125 with such servers Wget now simply ignores the data up to the byte
126 where the last attempt left off, and only then continues appending to
127 the file. That way the downloaded file never shrinks, and download
128 retries from servers without support for partial downloads work even
129 when downloading to stdout.
133 *** SSL/TLS downloads now attempt to verify the server's certificate
134 against the recognized certificate authorities. This requires CA
135 certificates to have been installed in a location visible to the
136 OpenSSL library. If this is not the case, you can get the bundle
137 yourself from a source you trust (for example, the bundle extracted
138 from Mozilla available at http://curl.haxx.se/docs/caextract.html),
139 and point Wget to the PEM file using the `--ca-certificate'
140 command-line option or the corresponding `.wgetrc' command.
142 *** Secure downloads now verify that the host name in the URL matches
143 the "common name" in the certificate presented by the server.
145 *** Although the above checks provide more secure downloads, they
146 unavoidably break interoperability with some sites that worked with
147 previous versions, particularly those using self-signed, expired, or
148 otherwise invalid certificates. If you encounter "certificate
149 verification" errors or complaints that "common name doesn't match
150 requested host name" and are convinced of the site's authenticity, you
151 can use `--no-check-certificate' to bypass both checks.
153 *** Talking to SSL/TLS servers over proxies now actually works.
154 Previous versions of Wget erroneously sent GET requests for https
155 URLs. Wget 1.10 utilizes the CONNECT method designed for this
158 *** The SSL/TLS-related options have been redesigned and, for the
159 first time, documented in the manual. The old, undocumented, options
160 are no longer supported.
162 ** Passive FTP is now the default FTP transfer mode. Use
163 `--no-passive-ftp' or specify `passive_ftp = off' in your init file to
164 revert to the old behavior.
166 ** The `--header' option can now be used to override generated
167 headers. For example, `wget --header="Host: foo.bar"
168 http://127.0.0.1' tells Wget to connect to localhost, but to specify
169 "foo.bar" in the `Host' header. In previous versions such use of
170 `--header' lead to duplicate headers in HTTP requests.
172 ** The responses without headers, aka "HTTP 0.9" responses, are
173 detected and handled. Although HTTP 0.9 has long been obsolete, it is
174 still occasionally used, sometimes by accident.
176 ** The progress bar is now updated regularly even when the data does
177 not arrive from the network.
179 ** Wget no longer preserves permissions of files retrieved by FTP by
180 default. Anonymous FTP servers frequently use permissions like "664",
181 which might not be what the user wants. The new option
182 `--preserve-permissions' and the corresponding `.wgetrc' variable can
183 be used to revert to the old behavior.
185 ** The new option `--protocol-directories' instructs Wget to also use
186 the protocol name as a directory component of local file names.
188 ** Options that previously unconditionally set or unset various flags
189 are now boolean options that can be invoked as either `--OPTION' or
190 `--no-OPTION'. Options that required an argument "on" or "off" have
191 also been changed this way, but they still accept the old syntax for
192 backward compatibility. For example, instead of `--glob=off' you can
195 Allowing `--no-OPTION' for every `--OPTION' and the other way around
196 is useful because it allows the user to override non-default behavior
197 specified via `.wgetrc'.
199 ** The new option `--keep-session-cookies' causes `--save-cookies' to
200 save session cookies (normally only kept in memory) along with the
201 permanent ones. This is useful because many sites track important
202 information, such as whether the user has authenticated, in session
203 cookies. With this option multiple Wget runs are treated as a single
206 ** Wget now supports the --ftp-user and --ftp-password command
207 switches to set username and password for FTP, and the --user and
208 --password command switches to set username and password for both FTP
209 and HTTP. The --http-passwd and --proxy-passwd command switches have
210 been renamed to --http-password and --proxy-password respectively, and
211 the related http_passwd and proxy_passwd .wgetrc commands to
212 http_password and proxy_password respectively. The login and passwd
213 .wgetrc commands have been deprecated.
215 * `wget -b' now works correctly under Windows.
217 * Wget 1.9.1 is a bugfix release with no user-visible changes.
219 * Changes in Wget 1.9.
221 ** It is now possible to specify that POST method be used for HTTP
222 requests. For example, `wget --post-data="id=foo&data=bar" URL' will
223 send a POST request with the specified contents.
225 ** IPv6 support is available, although it's still experimental.
227 ** The `--timeout' option now also affects DNS lookup and establishing
228 the TCP connection. Previously it only affected reading and writing
229 data. Those three timeouts can be set separately using
230 `--dns-timeout', `--connection-timeout', and `--read-timeout',
233 ** Download speed shown by the progress bar is based on the data
234 recently read, rather than the average speed of the entire download.
235 The ETA projection is still based on the overall average.
237 ** It is now possible to connect to FTP servers through FWTK
238 firewalls. Set ftp_proxy to an FTP URL, and Wget will automatically
239 log on to the proxy as "username@host".
241 ** The new option `--retry-connrefused' makes Wget retry downloads
242 even in the face of refused connections, which are otherwise
243 considered a fatal error.
245 ** The new option `--no-dns-cache' may be used to prevent Wget from
248 ** Wget no longer escapes characters in local file names based on
249 whether they're appropriate in URLs. Escaping can still occur for
250 nonprintable characters or for '/', but no longer for frequent
251 characters such as space. You can use the new option
252 --restrict-file-names to relax or strengthen these rules, which can be
253 useful if you dislike the default or if you're downloading to
254 non-native partitions.
256 ** Handling of HTML comments has been dumbed down to conform to what
257 users expect and other browsers do: instead of being treated as SGML
258 declaration, a comment is terminated at the first occurrence of "-->".
259 Use `--strict-comments' to revert to the old behavior.
261 ** Wget now correctly handles relative URIs that begin with "//", such
262 as "//img.foo.com/foo.jpg".
264 ** Boolean options in `.wgetrc' and on the command line now accept
265 values "yes" and "no" along with the traditional "on" and "off".
267 ** It is now possible to specify decimal values for timeouts, waiting
268 periods, and download rate. For instance, `--wait=0.5' now works as
269 expected, as does `--dns-timeout=0.5' and even `--limit-rate=2.5k'.
271 * Wget 1.8.2 is a bugfix release with no user-visible changes.
273 * Wget 1.8.1 is a bugfix release with no user-visible changes.
275 * Changes in Wget 1.8.
277 ** A new progress indicator is now available and used by default.
278 You can choose the progress bar type with `--progress=TYPE'. Two
279 types are available, "bar" (the new default), and "dot" (the old
280 dotted indicator). You can permanently revert to the old progress
281 indicator by putting `progress = dot' in your `.wgetrc'.
283 ** You can limit the download rate of the retrieval using the
284 `--limit-rate' option. For example, `wget --limit-rate=15k URL' will
285 tell Wget not to download the body of the URL faster than 15 kilobytes
288 ** Recursive retrieval and link conversion have been revamped:
290 *** Wget now traverses links breadth-first. This makes the
291 calculation of depth much more reliable than before. Also, recursive
292 downloads are faster and consume *significantly* less memory than
295 *** Links are converted only when the entire retrieval is complete.
296 This is the only safe thing to do, as only then is it known what URLs
297 have been downloaded.
299 *** BASE tags are handled correctly when converting links. Since Wget
300 already resolves <base href="..."> when resolving handling URLs, link
301 conversion now makes the BASE tags point to an empty string.
303 *** HTML anchors are now handled correctly. Links to an anchor in the
304 same document (<a href="#anchorname">), which used to confuse Wget,
305 are now converted correctly.
307 *** When in page-requisites (-p) mode, no-parent (-np) is ignored when
308 retrieving for inline images, stylesheets, and other documents needed
311 *** Page-requisites (-p) mode now works with frames. In other words,
312 `wget -p URL-THAT-USES-FRAMES' will now download the frame HTML files,
313 and all the files that they need to be displayed properly.
315 ** `--base' now works conjunction with `--input-file', providing a
316 base for each URL and thereby allowing the URLs in the file to be
319 ** If a host has more than one IP address, Wget uses the other
320 addresses when accessing the first one fails.
322 ** Host directories now contain port information if the URL is at a
325 ** Wget now supports the robots.txt directives specified in
326 <http://www.robotstxt.org/wc/norobots-rfc.txt>.
328 ** URL parser has been fixed, especially the infamous overzealous
329 quoting. Wget no longer dequotes reserved characters, e.g. `%3F' is
330 no longer translated to `?', nor `%2B' to `+'. Unsafe characters
331 which are not reserved are still escaped, of course.
333 ** No more than 20 successive redirections are allowed.
335 * Wget 1.7.1 is a bugfix release with no user-visible changes.
337 * Changes in Wget 1.7.
339 ** SSL (`https') pages now work if you compile Wget with SSL support;
340 use the `--with-ssl' configure flag. You need to have OpenSSL
343 ** Cookies are now supported. Wget will accept cookies sent by the
344 server and return them in later requests. Additionally, it can load
345 and save cookies to disk, in the same format that Netscape uses.
347 ** "Keep-alive" (persistent) HTTP connections are now supported.
348 Using keep-alive allows Wget to share one TCP/IP connection for
349 many retrievals, making multiple-file downloads faster and less
350 stressing for the server and the network.
352 ** Wget now recognizes FTP directory listings generated by NT and VMS
355 ** It is now possible to recurse through FTP sites where logging in
356 puts you in some directory other than '/'.
358 ** You may now use `~' to mean home directory in `.wgetrc'. For
359 example, `load_cookies = ~/.netscape/cookies.txt' works as you would
362 ** The HTML parser has been rewritten. The new one works more
363 reliably, allows finer-grained control over which tags and attributes
364 are detected, and has better support for some features like correctly
365 skipping comments and declarations, decoding entities, etc. It is
368 ** <meta name="robots"> tags are now respected.
370 ** Wget's internal tables now use hash tables instead of linked lists
371 where appropriate. This results in huge speedups when retrieving
372 large sites (thousands of documents).
374 ** Wget now has a man page, automatically generated from the Texinfo
375 documentation. (The last version that shipped with a man page was
376 1.4.5). To get this, you need to have pod2man from the Perl
377 distribution installed on your system.
379 * Changes in Wget 1.6
381 ** Administrative changes.
383 *** Maintainership. Due to Hrvoje being plagued with a "real job",
384 Dan Harkless is the most active maintainer (not that he doesn't have a
385 real job as well). Hrvoje still participates occasionally, and both
386 are being helped by many other people.
388 *** Web page. Thanks to Jan Prikryl, Wget has an "official" web page.
391 http://sunsite.dk/wget/
393 *** Anonymous CVS. Thanks to ever-helpful Karsten Thygesen, Wget
394 sources are now available at an anonymous CVS server. Take a look at
395 the web page for downloading instructions.
397 ** New -K / --backup-converted / backup_converted = on option causes files
398 modified due to -k to be saved with a .orig prefix before being changed. When
399 using -N as well, it is these .orig files that are compared against the server.
401 ** New --follow-tags / follow_tags = ... option allows you to restrict
402 Wget to following only certain HTML tags when doing a recursive
403 retrieval. -G / --ignore-tags / ignore_tags = ... is just the
404 opposite -- all tags but the ones you specify will be followed.
406 ** New --waitretry / waitretry = SECONDS option allows waiting between retries
407 of failed downloads. Wget will use "linear" backoff, waiting 1 second after the
408 first failure, 2 after the second, up to SECONDS. waitretry is set to 10 by
409 default in the system wgetrc.
411 ** New -p / --page-requisites / page_requisites = on option causes
412 Wget to download all ancillary files necessary to display a given HTML
413 page properly (e.g. inlined images).
415 ** New -E / --html-extension / html_extension = on option causes Wget
416 to append ".html" to text/html filenames not ending in regexp
417 "\.[Hh][Tt][Mm][Ll]?".
419 ** New type of .wgetrc command -- "lockable Boolean". Can be set to on, off,
420 always, or never. This allows the .wgetrc to override the commandline. So far,
421 passive_ftp is the only .wgetrc command which takes a lockable Boolean.
423 ** A number of new translation files have been added.
425 ** New --bind-address / bind_address = <address> option for people on hosts
426 bound to multiple IP addresses.
428 ** wget now accepts (illegal per HTTP spec) relative URLs in HTTP redirects.
430 * Wget 1.5.3 is a bugfix release with no user-visible changes.
432 * Wget 1.5.2 is a bugfix release with no user-visible changes.
434 * Wget 1.5.1 is a bugfix release with no user-visible changes.
436 * Changes in Wget 1.5.0
438 ** Wget speaks many languages!
440 On systems with gettext(), Wget will output messages in the language
441 set by the current locale, if available. At this time we support
442 Czech, German, Croatian, Italian, Norwegian and Portuguese.
444 ** Opie (Skey) is now supported with FTP.
446 ** HTTP Digest Access Authentication (RFC2069) is now supported.
448 ** The new `-b' option makes Wget go to background automatically.
450 ** The `-I' and `-X' options now accept wildcard arguments.
452 ** The `-w' option now accepts suffixes `s' for seconds, `m' for
453 minutes, `h' for hours, `d' for days and `w' for weeks.
455 ** Upon getting SIGHUP, the whole previous log is now copied to
458 ** Wget now understands proxy settings with explicit usernames and
459 passwords, e.g. `http://user:password@proxy.foo.com/'.
461 ** You can use the new `--cut-dirs' option to make Wget create less
464 ** The `;type=a' appendix to FTP URLs is now recognized. For
465 instance, the following command will retrieve the welcoming message in
468 wget "ftp://ftp.somewhere.com/welcome.msg;type=a"
470 ** `--help' and `--version' options have been redone to to conform to
471 standards set by other GNU utilities.
473 ** Wget should now be compilable under MS Windows environment. MS
474 Visual C++ and Watcom C have been used successfully.
476 ** If the file length is known, percentages are displayed during
479 ** The manual page, now hopelessly out of date, is no longer
480 distributed with Wget.
482 * Wget 1.4.5 is a bugfix release with no user-visible changes.
484 * Wget 1.4.4 is a bugfix release with no user-visible changes.
486 * Changes in Wget 1.4.3
488 ** Wget is now a GNU utility.
490 ** Can do passive FTP.
494 ** Info documentation expanded.
496 ** Compiles on pre-ANSI compilers.
498 ** Global wgetrc now goes to /usr/local/etc (i.e. $sysconfdir).
502 * Changes in Wget 1.4.2
504 ** New mirror site at ftp://sunsite.auc.dk/pub/infosystems/wget/,
505 thanks to Karsten Thygesen.
507 ** Mailing list! Mail to wget-request@sunsite.auc.dk to subscribe.
509 ** New option --delete-after for proxy prefetching.
511 ** New option --retr-symlinks to retrieve symbolic links like plain
514 ** rmold.pl -- script to remove files deleted on the remote server
516 ** --convert-links should work now.
520 * Changes in Wget 1.4.1
524 ** Added -I (the opposite of -X).
526 ** Dot tracing is now customizable; try wget --dot-style=binary
528 * Changes in Wget 1.4.0
530 ** Wget 1.4.0 [formerly known as Geturl] is an extensive rewrite of
531 Geturl. Although many things look suspiciously similar, most of the
532 stuff was rewritten, like recursive retrieval, HTTP, FTP and mostly
533 everything else. Wget should be now easier to debug, maintain and,
534 most importantly, use.
536 ** Recursive HTTP should now work without glitches, even with Location
537 changes, server-generated directory listings and other naughty stuff.
539 ** HTTP regetting is supported on servers that support Range
540 specification. WWW authorization is supported -- try
541 wget http://user:password@hostname/
543 ** FTP support was rewritten and widely enhanced. Globbing should now
544 work flawlessly. Symbolic links are created locally. All the
545 information the Unix-style ls listing can give is now recognized.
547 ** Recursive FTP is supported, e.g.
548 wget -r ftp://gnjilux.cc.fer.hr/pub/unix/util/
550 ** You can specify "rejected" directories, to which you do not want to
551 enter, e.g. with wget -X /pub
553 ** Time-stamping is supported, with both HTTP and FTP. Try wget -N URL.
555 ** A new texinfo reference manual is provided. It can be read with
556 Emacs, standalone info, or converted to HTML, dvi or postscript.
558 ** Fixed a long-standing bug, so that Wget now works over SLIP
561 ** You can have a system-wide wgetrc (/usr/local/lib/wgetrc by
562 default). Settings in $HOME/.wgetrc override the global ones, of
565 ** You can set up quota in .wgetrc to prevent sucking too much
566 data. Try `quota = 5M' in .wgetrc (or quota = 100K if you want your
567 sysadmin to like you).
569 ** Download rate is printed after retrieval.
571 ** Wget now sends the `Referer' header when retrieving
574 ** With the new --no-parent option Wget can retrieve FTP recursively
575 through a proxy server.
577 ** HTML parser, as well as the whole of Wget was rewritten to be much
578 faster and less memory-consuming (yes, both).
580 ** Absolute links can be converted to relative links locally. Check
583 ** Wget catches hangup, filtering the output to a log file and
584 resuming work. Try kill -HUP %?wget.
586 ** User-defined headers can be sent. Try
588 wget http://fly.cc.her.hr/ --header='Accept-Charset: iso-8859-2'
590 ** Acceptance/Rejection lists may contain wildcards.
592 ** Wget can display HTTP headers and/or FTP server response with the
593 new `-S' option. It can save the original HTTP headers with `-s'.
595 ** socks library is now supported (thanks to Antonio Rosella
596 <Antonio.Rosella@agip.it>). Configure with --with-socks.
598 ** There is a nicer display of REST-ed output.
600 ** Many new options (like -x to force directory hierarchy, or -m to
601 turn on mirroring options).
603 ** Wget is now distributed under GNU General Public License (GPL).
605 ** Lots of small features I can't remember. :-)
607 ** A host of bugfixes.
609 * Changes in Geturl 1.3
611 ** Added FTP globbing support (ftp://fly.cc.fer.hr/*)
613 ** Added support for no_proxy
615 ** Added support for ftp://user:password@host/
617 ** Added support for %xx in URL syntax
619 ** More natural command-line options
621 ** Added -e switch to execute .geturlrc commands from the command-line
623 ** Added support for robots.txt
625 ** Fixed some minor bugs
627 * Geturl 1.2 is a bugfix release with no user-visible changes.
629 * Changes in Geturl 1.1
631 ** REST supported in FTP
633 ** Proxy servers supported
635 ** GNU getopt used, which enables command-line arguments to be ordered
636 as you wish, e.g. geturl http://fly.cc.fer.hr/ -vo log is the same as
637 geturl -vo log http://fly.cc.fer.hr/
639 ** Netscape-compatible URL syntax for HTTP supported: host[:port]/dir/file
641 ** NcFTP-compatible colon URL syntax for FTP supported: host:/dir/file
643 ** <base href="xxx"> supported
645 ** autoconf supported
647 ----------------------------------------------------------------------
648 Copyright information:
650 Copyright (C) 1997-2005 Free Software Foundation, Inc.
652 Permission is granted to anyone to make or distribute verbatim
653 copies of this document as received, in any medium, provided that
654 the copyright notice and this permission notice are preserved, thus
655 giving the recipient permission to redistribute in turn.
657 Permission is granted to distribute modified versions of this
658 document, or of portions of it, under the above conditions,
659 provided also that they carry prominent notices stating who last