1URI(7) Linux Programmer's Manual URI(7)
2
6 uri, url, urn - uniform resource identifier (URI), including a URL or
7 URN
8
10 URI = [ absoluteURI | relativeURI ] [ "#" fragment ]
11 absoluteURI = scheme ":" ( hierarchical_part | opaque_part )
13 relativeURI = ( net_path | absolute_path | relative_path ) [ "?" query ]
15 scheme = "http" | "ftp" | "gopher" | "mailto" | "news" | "telnet" |
17 "file" | "man" | "info" | "whatis" | "ldap" | "wais" | ...
18 hierarchical_part = ( net_path | absolute_path ) [ "?" query ]
20 net_path = "//" authority [ absolute_path ]
22 absolute_path = "/" path_segments
24 relative_path = relative_segment [ absolute_path ]
26
28 A Uniform Resource Identifier (URI) is a short string of characters
29 identifying an abstract or physical resource (for example, a web page).
30 A Uniform Resource Locator (URL) is a URI that identifies a resource
31 through its primary access mechanism (e.g., its network "location"),
32 rather than by name or some other attribute of that resource. A Uni‐
33 form Resource Name (URN) is a URI that must remain globally unique and
34 persistent even when the resource ceases to exist or becomes unavail‐
35 able.
36 URIs are the standard way to name hypertext link destinations for tools
38 such as web browsers. The string "http://www.kernel.org" is a URL (and
39 thus it is also a URI). Many people use the term URL loosely as a syn‐
40 onym for URI (though technically URLs are a subset of URIs).
41 URIs can be absolute or relative. An absolute identifier refers to a
43 resource independent of context, while a relative identifier refers to
44 a resource by describing the difference from the current context.
45 Within a relative path reference, the complete path segments "." and
46 ".." have special meanings: "the current hierarchy level" and "the
47 level above this hierarchy level", respectively, just like they do in
48 UNIX-like systems. A path segment which contains a colon character
49 can't be used as the first segment of a relative URI path (e.g.,
50 "this:that"), because it would be mistaken for a scheme name; precede
51 such segments with ./ (e.g., "./this:that"). Note that descendants of
52 MS-DOS (e.g., Microsoft Windows) replace devicename colons with the
53 vertical bar ("|") in URIs, so "C:" becomes "C|".
54 A fragment identifier, if included, refers to a particular named por‐
56 tion (fragment) of a resource; text after a '#' identifies the frag‐
57 ment. A URI beginning with '#' refers to that fragment in the current
58 resource.
59 Usage
61 There are many different URI schemes, each with specific additional
62 rules and meanings, but they are intentionally made to be as similar as
63 possible. For example, many URL schemes permit the authority to be the
64 following format, called here an ip_server (square brackets show what's
65 optional):
66 ip_server = [user [ : password ] @ ] host [ : port]
68 This format allows you to optionally insert a username, a user plus
70 password, and/or a port number. The host is the name of the host com‐
71 puter, either its name as determined by DNS or an IP address (numbers
72 separated by periods). Thus the URI <http://fred:fredpassword@exam‐
73 ple.com:8080/> logs into a web server on host example.com as fred
74 (using fredpassword) using port 8080. Avoid including a password in a
75 URI if possible because of the many security risks of having a password
76 written down. If the URL supplies a username but no password, and the
77 remote server requests a password, the program interpreting the URL
78 should request one from the user.
79 Here are some of the most common schemes in use on UNIX-like systems
81 that are understood by many tools. Note that many tools using URIs
82 also have internal schemes or specialized schemes; see those tools'
83 documentation for information on those schemes.
84 http - Web (HTTP) server
86 http://ip_server/path
88 http://ip_server/path?query
89 This is a URL accessing a web (HTTP) server. The default port is 80.
91 If the path refers to a directory, the web server will choose what to
92 return; usually if there is a file named "index.html" or "index.htm"
93 its content is returned, otherwise, a list of the files in the current
94 directory (with appropriate links) is generated and returned. An exam‐
95 ple is <http://lwn.net>.
96 A query can be given in the archaic "isindex" format, consisting of a
98 word or phrase and not including an equal sign (=). A query can also
99 be in the longer "GET" format, which has one or more query entries of
100 the form key=value separated by the ampersand character (&). Note that
101 key can be repeated more than once, though it's up to the web server
102 and its application programs to determine if there's any meaning to
103 that. There is an unfortunate interaction with HTML/XML/SGML and the
104 GET query format; when such URIs with more than one key are embedded in
105 SGML/XML documents (including HTML), the ampersand (&) has to be
106 rewritten as &. Note that not all queries use this format; larger
107 forms may be too long to store as a URI, so they use a different inter‐
108 action mechanism (called POST) which does not include the data in the
109 URI. See the Common Gateway Interface specification at
110 ⟨http://www.w3.org/CGI⟩ for more information.
111 ftp - File Transfer Protocol (FTP)
113 ftp://ip_server/path
115 This is a URL accessing a file through the file transfer protocol
117 (FTP). The default port (for control) is 21. If no username is
118 included, the username "anonymous" is supplied, and in that case many
119 clients provide as the password the requestor's Internet email address.
120 An example is <ftp://ftp.is.co.za/rfc/rfc1808.txt>.
121 gopher - Gopher server
123 gopher://ip_server/gophertype selector
125 gopher://ip_server/gophertype selector%09search
126 gopher://ip_server/gophertype selector%09search%09gopher+_string
127 The default gopher port is 70. gophertype is a single-character field
129 to denote the Gopher type of the resource to which the URL refers. The
130 entire path may also be empty, in which case the delimiting "/" is also
131 optional and the gophertype defaults to "1".
132 selector is the Gopher selector string. In the Gopher protocol, Gopher
134 selector strings are a sequence of octets which may contain any octets
135 except 09 hexadecimal (US-ASCII HT or tab), 0A hexadecimal (US-ASCII
136 character LF), and 0D (US-ASCII character CR).
137 mailto - Email address
139 mailto:email-address
141 This is an email address, usually of the form name@hostname. See
143 mailaddr(7) for more information on the correct format of an email
144 address. Note that any % character must be rewritten as %25. An exam‐
145 ple is <mailto:dwheeler@dwheeler.com>.
146 news - Newsgroup or News message
148 news:newsgroup-name
150 news:message-id
151 A newsgroup-name is a period-delimited hierarchical name, such as
153 "comp.infosystems.www.misc". If <newsgroup-name> is "*" (as in
154 <news:*>), it is used to refer to "all available news groups". An
155 example is <news:comp.lang.ada>.
156 A message-id corresponds to the Message-ID of IETF RFC 1036,
158 ⟨http://www.ietf.org/rfc/rfc1036.txt⟩ without the enclosing "<" and
159 ">"; it takes the form unique@full_domain_name. A message identifier
160 may be distinguished from a news group name by the presence of the "@"
161 character.
162 telnet - Telnet login
164 telnet://ip_server/
166 The Telnet URL scheme is used to designate interactive text services
168 that may be accessed by the Telnet protocol. The final "/" character
169 may be omitted. The default port is 23. An example is <tel‐
170 net://melvyl.ucop.edu/>.
171 file - Normal file
173 file://ip_server/path_segments
175 file:path_segments
176 This represents a file or directory accessible locally. As a special
178 case, ip_server can be the string "localhost" or the empty string; this
179 is interpreted as "the machine from which the URL is being inter‐
180 preted". If the path is to a directory, the viewer should display the
181 directory's contents with links to each containee; not all viewers cur‐
182 rently do this. KDE supports generated files through the URL
183 <file:/cgi-bin>. If the given file isn't found, browser writers may
184 want to try to expand the filename via filename globbing (see glob(7)
185 and glob(3)).
186 The second format (e.g., <file:/etc/passwd>) is a correct format for
188 referring to a local file. However, older standards did not permit
189 this format, and some programs don't recognize this as a URI. A more
190 portable syntax is to use an empty string as the server name, for exam‐
191 ple, <file:///etc/passwd>; this form does the same thing and is easily
192 recognized by pattern matchers and older programs as a URI. Note that
193 if you really mean to say "start from the current location," don't
194 specify the scheme at all; use a relative address like <../test.txt>,
195 which has the side-effect of being scheme-independent. An example of
196 this scheme is <file:///etc/passwd>.
197 man - Man page documentation
199 man:command-name
201 man:command-name(section)
202 This refers to local online manual (man) reference pages. The command
204 name can optionally be followed by a parenthesis and section number;
205 see man(7) for more information on the meaning of the section numbers.
206 This URI scheme is unique to UNIX-like systems (such as Linux) and is
207 not currently registered by the IETF. An example is <man:ls(1)>.
208 info - Info page documentation
210 info:virtual-filename
212 info:virtual-filename#nodename
213 info:(virtual-filename)
214 info:(virtual-filename)nodename
215 This scheme refers to online info reference pages (generated from tex‐
217 info files), a documentation format used by programs such as the GNU
218 tools. This URI scheme is unique to UNIX-like systems (such as Linux)
219 and is not currently registered by the IETF. As of this writing, GNOME
220 and KDE differ in their URI syntax and do not accept the other's syn‐
221 tax. The first two formats are the GNOME format; in nodenames all spa‐
222 ces are written as underscores. The second two formats are the KDE
223 format; spaces in nodenames must be written as spaces, even though this
224 is forbidden by the URI standards. It's hoped that in the future most
225 tools will understand all of these formats and will always accept
226 underscores for spaces in nodenames. In both GNOME and KDE, if the
227 form without the nodename is used the nodename is assumed to be "Top".
228 Examples of the GNOME format are <info:gcc> and <info:gcc#G++_and_GCC>.
229 Examples of the KDE format are <info:(gcc)> and <info:(gcc)G++ and
230 GCC>.
231 whatis - Documentation search
233 whatis:string
235 This scheme searches the database of short (one-line) descriptions of
237 commands and returns a list of descriptions containing that string.
238 Only complete word matches are returned. See whatis(1). This URI
239 scheme is unique to UNIX-like systems (such as Linux) and is not cur‐
240 rently registered by the IETF.
241 ghelp - GNOME help documentation
243 ghelp:name-of-application
245 This loads GNOME help for the given application. Note that not much
247 documentation currently exists in this format.
248 ldap - Lightweight Directory Access Protocol
250 ldap://hostport
252 ldap://hostport/
253 ldap://hostport/dn
254 ldap://hostport/dn?attributes
255 ldap://hostport/dn?attributes?scope
256 ldap://hostport/dn?attributes?scope?filter
257 ldap://hostport/dn?attributes?scope?filter?extensions
258 This scheme supports queries to the Lightweight Directory Access Proto‐
260 col (LDAP), a protocol for querying a set of servers for hierarchically
261 organized information (such as people and computing resources). See
262 RFC 2255 ⟨http://www.ietf.org/rfc/rfc2255.txt⟩ for more information on
263 the LDAP URL scheme. The components of this URL are:
264 hostport the LDAP server to query, written as a hostname optionally
266 followed by a colon and the port number. The default LDAP
267 port is TCP port 389. If empty, the client determines
268 which the LDAP server to use.
269 dn the LDAP Distinguished Name, which identifies the base
271 object of the LDAP search (see RFC 2253
272 ⟨http://www.ietf.org/rfc/rfc2253.txt⟩ section 3).
273 attributes a comma-separated list of attributes to be returned; see
275 RFC 2251 section 4.1.5. If omitted, all attributes should
276 be returned.
277 scope specifies the scope of the search, which can be one of
279 "base" (for a base object search), "one" (for a one-level
280 search), or "sub" (for a subtree search). If scope is
281 omitted, "base" is assumed.
282 filter specifies the search filter (subset of entries to return).
284 If omitted, all entries should be returned. See RFC 2254
285 ⟨http://www.ietf.org/rfc/rfc2254.txt⟩ section 4.
286 extensions a comma-separated list of type=value pairs, where the
288 =value portion may be omitted for options not requiring it.
289 An extension prefixed with a '!' is critical (must be sup‐
290 ported to be valid), otherwise it is noncritical
291 (optional).
292 LDAP queries are easiest to explain by example. Here's a query that
294 asks ldap.itd.umich.edu for information about the University of Michi‐
295 gan in the U.S.:
296 ldap://ldap.itd.umich.edu/o=University%20of%20Michigan,c=US
298 To just get its postal address attribute, request:
300 ldap://ldap.itd.umich.edu/o=University%20of%20Michigan,c=US?postalAddress
302 To ask a host.com at port 6666 for information about the person with
304 common name (cn) "Babs Jensen" at University of Michigan, request:
305 ldap://host.com:6666/o=University%20of%20Michigan,c=US??sub?(cn=Babs%20Jensen)
307 wais - Wide Area Information Servers
309 wais://hostport/database
311 wais://hostport/database?search
312 wais://hostport/database/wtype/wpath
313 This scheme designates a WAIS database, search, or document (see IETF
315 RFC 1625 ⟨http://www.ietf.org/rfc/rfc1625.txt⟩ for more information on
316 WAIS). Hostport is the hostname, optionally followed by a colon and
317 port number (the default port number is 210).
318 The first form designates a WAIS database for searching. The second
320 form designates a particular search of the WAIS database database. The
321 third form designates a particular document within a WAIS database to
322 be retrieved. wtype is the WAIS designation of the type of the object
323 and wpath is the WAIS document-id.
324 other schemes
326 There are many other URI schemes. Most tools that accept URIs support
328 a set of internal URIs (e.g., Mozilla has the about: scheme for inter‐
329 nal information, and the GNOME help browser has the toc: scheme for
330 various starting locations). There are many schemes that have been
331 defined but are not as widely used at the current time (e.g., pros‐
332 pero). The nntp: scheme is deprecated in favor of the news: scheme.
333 URNs are to be supported by the urn: scheme, with a hierarchical name
334 space (e.g., urn:ietf:... would identify IETF documents); at this time
335 URNs are not widely implemented. Not all tools support all schemes.
336 Character encoding
338 URIs use a limited number of characters so that they can be typed in
339 and used in a variety of situations.
340 The following characters are reserved, that is, they may appear in a
342 URI but their use is limited to their reserved purpose (conflicting
343 data must be escaped before forming the URI):
344 ; / ? : @ & = + $ ,
346 Unreserved characters may be included in a URI. Unreserved characters
348 include uppercase and lowercase English letters, decimal digits, and
349 the following limited set of punctuation marks and symbols:
350 - _ . ! ~ * ' ( )
352 All other characters must be escaped. An escaped octet is encoded as a
354 character triplet, consisting of the percent character "%" followed by
355 the two hexadecimal digits representing the octet code (you can use
356 uppercase or lowercase letters for the hexadecimal digits). For exam‐
357 ple, a blank space must be escaped as "%20", a tab character as "%09",
358 and the "&" as "%26". Because the percent "%" character always has the
359 reserved purpose of being the escape indicator, it must be escaped as
360 "%25". It is common practice to escape space characters as the plus
361 symbol (+) in query text; this practice isn't uniformly defined in the
362 relevant RFCs (which recommend %20 instead) but any tool accepting URIs
363 with query text should be prepared for them. A URI is always shown in
364 its "escaped" form.
365 Unreserved characters can be escaped without changing the semantics of
367 the URI, but this should not be done unless the URI is being used in a
368 context that does not allow the unescaped character to appear. For
369 example, "%7e" is sometimes used instead of "~" in an HTTP URL path,
370 but the two are equivalent for an HTTP URL.
371 For URIs which must handle characters outside the US ASCII character
373 set, the HTML 4.01 specification (section B.2) and IETF RFC 2718 (sec‐
374 tion 2.2.5) recommend the following approach:
375 1. translate the character sequences into UTF-8 (IETF RFC 2279)—see
377 utf-8(7)—and then
378 2. use the URI escaping mechanism, that is, use the %HH encoding for
380 unsafe octets.
381 Writing a URI
383 When written, URIs should be placed inside double quotes (e.g.,
384 "http://www.kernel.org"), enclosed in angle brackets (e.g.,
385 <http://lwn.net>), or placed on a line by themselves. A warning for
386 those who use double-quotes: never move extraneous punctuation (such as
387 the period ending a sentence or the comma in a list) inside a URI,
388 since this will change the value of the URI. Instead, use angle brack‐
389 ets instead, or switch to a quoting system that never includes extrane‐
390 ous characters inside quotation marks. This latter system, called the
391 'new' or 'logical' quoting system by "Hart's Rules" and the "Oxford
392 Dictionary for Writers and Editors", is preferred practice in Great
393 Britain and hackers worldwide (see the Jargon File's section on Hacker
394 Writing Style, ⟨http://www.fwi.uva.nl/~mes/jargon/h
395 /HackerWritingStyle.html⟩, for more information). Older documents sug‐
396 gested inserting the prefix "URL:" just before the URI, but this form
397 has never caught on.
398 The URI syntax was designed to be unambiguous. However, as URIs have
400 become commonplace, traditional media (television, radio, newspapers,
401 billboards, etc.) have increasingly used abbreviated URI references
402 consisting of only the authority and path portions of the identified
403 resource (e.g., <www.w3.org/Addressing>). Such references are primar‐
404 ily intended for human interpretation rather than machine, with the
405 assumption that context-based heuristics are sufficient to complete the
406 URI (e.g., hostnames beginning with "www" are likely to have a URI pre‐
407 fix of "http://" and hostnames beginning with "ftp" likely to have a
408 prefix of "ftp://"). Many client implementations heuristically resolve
409 these references. Such heuristics may change over time, particularly
410 when new schemes are introduced. Since an abbreviated URI has the same
411 syntax as a relative URL path, abbreviated URI references cannot be
412 used where relative URIs are permitted, and can be used only when there
413 is no defined base (such as in dialog boxes). Don't use abbreviated
414 URIs as hypertext links inside a document; use the standard format as
415 described here.
416
418 (IETF RFC 2396) ⟨http://www.ietf.org/rfc/rfc2396.txt⟩, (HTML 4.0)
419 ⟨http://www.w3.org/TR/REC-html40⟩.
420
422 Any tool accepting URIs (e.g., a web browser) on a Linux system should
423 be able to handle (directly or indirectly) all of the schemes described
424 here, including the man: and info: schemes. Handling them by invoking
425 some other program is fine and in fact encouraged.
426 Technically the fragment isn't part of the URI.
428 For information on how to embed URIs (including URLs) in a data format,
430 see documentation on that format. HTML uses the format <A HREF="uri">
431 text </A>. Texinfo files use the format @uref{uri}. Man and mdoc have
432 the recently added UR macro, or just include the URI in the text (view‐
433 ers should be able to detect :// as part of a URI).
434 The GNOME and KDE desktop environments currently vary in the URIs they
436 accept, in particular in their respective help browsers. To list man
437 pages, GNOME uses <toc:man> while KDE uses <man:(index)>, and to list
438 info pages, GNOME uses <toc:info> while KDE uses <info:(dir)> (the
439 author of this man page prefers the KDE approach here, though a more
440 regular format would be even better). In general, KDE uses <file:/cgi-
441 bin/> as a prefix to a set of generated files. KDE prefers documenta‐
442 tion in HTML, accessed via the <file:/cgi-bin/helpindex>. GNOME
443 prefers the ghelp scheme to store and find documentation. Neither
444 browser handles file: references to directories at the time of this
445 writing, making it difficult to refer to an entire directory with a
446 browsable URI. As noted above, these environments differ in how they
447 handle the info: scheme, probably the most important variation. It is
448 expected that GNOME and KDE will converge to common URI formats, and a
449 future version of this man page will describe the converged result.
450 Efforts to aid this convergence are encouraged.
451 Security
453 A URI does not in itself pose a security threat. There is no general
454 guarantee that a URL, which at one time located a given resource, will
455 continue to do so. Nor is there any guarantee that a URL will not
456 locate a different resource at some later point in time; such a guaran‐
457 tee can be obtained only from the person(s) controlling that namespace
458 and the resource in question.
459 It is sometimes possible to construct a URL such that an attempt to
461 perform a seemingly harmless operation, such as the retrieval of an
462 entity associated with the resource, will in fact cause a possibly dam‐
463 aging remote operation to occur. The unsafe URL is typically con‐
464 structed by specifying a port number other than that reserved for the
465 network protocol in question. The client unwittingly contacts a site
466 that is in fact running a different protocol. The content of the URL
467 contains instructions that, when interpreted according to this other
468 protocol, cause an unexpected operation. An example has been the use
469 of a gopher URL to cause an unintended or impersonating message to be
470 sent via a SMTP server.
471 Caution should be used when using any URL that specifies a port number
473 other than the default for the protocol, especially when it is a number
474 within the reserved space.
475 Care should be taken when a URI contains escaped delimiters for a given
477 protocol (for example, CR and LF characters for telnet protocols) that
478 these are not unescaped before transmission. This might violate the
479 protocol, but avoids the potential for such characters to be used to
480 simulate an extra operation or parameter in that protocol, which might
481 lead to an unexpected and possibly harmful remote operation to be per‐
482 formed.
483 It is clearly unwise to use a URI that contains a password which is
485 intended to be secret. In particular, the use of a password within the
486 "userinfo" component of a URI is strongly recommended against except in
487 those rare cases where the "password" parameter is intended to be pub‐
488 lic.
489
491 Documentation may be placed in a variety of locations, so there cur‐
492 rently isn't a good URI scheme for general online documentation in
493 arbitrary formats. References of the form <file:///usr/doc/ZZZ> don't
494 work because different distributions and local installation require‐
495 ments may place the files in different directories (it may be in
496 /usr/doc, or /usr/local/doc, or /usr/share, or somewhere else). Also,
497 the directory ZZZ usually changes when a version changes (though file‐
498 name globbing could partially overcome this). Finally, using the file:
499 scheme doesn't easily support people who dynamically load documentation
500 from the Internet (instead of loading the files onto a local filesys‐
501 tem). A future URI scheme may be added (e.g., "userdoc:") to permit
502 programs to include cross-references to more detailed documentation
503 without having to know the exact location of that documentation.
504 Alternatively, a future version of the filesystem specification may
505 specify file locations sufficiently so that the file: scheme will be
506 able to locate documentation.
507 Many programs and file formats don't include a way to incorporate or
509 implement links using URIs.
510 Many programs can't handle all of these different URI formats; there
512 should be a standard mechanism to load an arbitrary URI that automati‐
513 cally detects the users' environment (e.g., text or graphics, desktop
514 environment, local user preferences, and currently executing tools) and
515 invokes the right tool for any URI.
516
518 lynx(1), man2html(1), mailaddr(7), utf-8(7)
519 IETF RFC 2255 ⟨http://www.ietf.org/rfc/rfc2255.txt⟩
521
523 This page is part of release 5.04 of the Linux man-pages project. A
524 description of the project, information about reporting bugs, and the
525 latest version of this page, can be found at
526 https://www.kernel.org/doc/man-pages/.
527Linux 2017-09-15 URI(7)