W3C-URIS(2) W3C-URIS(2)
NAME
w3c-uris - uniform resource identifiers
SYNOPSIS
include "uris.m";
uris := load URIs URIs->PATH;
URI: import uris;
URI: adt
{
scheme: string;
userinfo: string; # authority, part I
host: string; # authority, part II
port: string; # authority, part III
path: string; # starts with / if absolute
query: string; # includes ? if not nil
fragment: string; # includes # if not nil
parse: fn(s: string): ref URI;
text: fn(u: self ref URI): string;
addbase: fn(base: self ref URI, rel: ref URI): ref URI;
authority: fn(u: self ref URI): string;
copy: fn(u: self ref URI): ref URI;
eq: fn(u: self ref URI, v: ref URI): int;
hasauthority: fn(u: self ref URI): int;
isabsolute: fn(u: self ref URI): int;
nodots: fn(u: self ref URI): ref URI;
userpw: fn(u: self ref URI): (string, string);
};
init: fn();
dec: fn(s: string): string;
enc: fn(s: string, safe: string): string;
DESCRIPTION
URIs supports the `generic syntax' for `Uniform' Resource
Identifiers (URIs), defined by RFC3986. Each URI can have
up to five components in the general syntax:
scheme: //authority/path ?query #fragment
where each component is optional, and can have scheme-
specific substructure. For instance, in the ftp, http
schemes, and perhaps others, the authority component has the
further syntax:
userinfo@host:port
The set of characters allowed in most components is also
Page 1 Plan 9 (printed 10/29/25)
W3C-URIS(2) W3C-URIS(2)
scheme-specific, as is their interpretation, and indeed the
interpretation of the component itself.
Init must be called before any other operation in the mod-
ule.
URI represents a parse of a URI into its components, where
the authority has been further split into the scheme-
specific but common triple of userinfo, host and port. (The
function URI.authority will reproduce the original authority
component if required.) The query field starts with the `?'
character that introduces the query component, so that an
empty query is represented by the string "?", and the
absence of a query component is represented by a nil value.
The fragment field is handled in a similar way with its
delimiting `#'. The fields representing the other compo-
nents do not include the delimiters in the syntax, and all
but query have percent-encoded characters decoded. (The
query string is an exception because the set of characters
to escape is application-specific. See below for decoding
and encoding functions.) URI provides the following opera-
tions:
parse(s)
Return a URI value representing the results of parsing
string s as a URI. There is no error return. The com-
ponent values have percent-escapes decoded as discussed
above. The scheme name is converted to lower case.
u.text()
Return the textual representation of u in the generic
syntax, adding percent-encoding as required to prevent
characters being misinterpreted as delimiters.
u.addbase(b)
Resolves URI reference u with respect to a base URI b,
including resolving all `.' and `..' segments in the
URI's path, and returns the resulting URI value. If u
is an absolute URI reference or b is nil, the result is
the same as u except that all `.' and `..' segments
have been resolved in the resulting path, and leading
instances of them removed.
u.authority()
Returns the text of the authority component of u, in
the generic syntax, created from its userinfo, host and
port components.
u.copy()
Return a reference to an independent copy of u.
u.eq(v)
Page 2 Plan 9 (printed 10/29/25)
W3C-URIS(2) W3C-URIS(2)
Returns true if u and v are textually equal in all com-
ponents except fragment. Note that u and v are assumed
to be in a canonical form for the scheme and applica-
tion.
u.eqf(v)
Returns true if u and v are textually equal in all com-
ponents including fragment.
u.hasauthority()
Returns true if any of the authority subcomponents of u
are not nil; returns false otherwise.
u.isabsolute()
Returns true if u has a scheme component; returns false
otherwise.
u.nodots()
Returns a new URI value in which all `.' and `..'
segments have been resolved (equivalent to
u.addbase(nil)).
u.userpw()
Returns a tuple (username, password) derived from pars-
ing the userinfo subcomponent of authority using the
deprecated but depressingly still common convention
that userinfo has the syntax ``username:password''.
A reserved or otherwise special character that appears in a
URI component must be encoded using a sequence of one or
more strings of the form %xx where xx is the hexadecimal
value of one byte of the character. A string s containing
such encodings can be decoded by the function dec. A string
s can be encoded by enc, where the parameter safe lists the
characters that need not be escaped (where safe may be nil
or empty). These functions are normally only needed to
decode and encode the values of URI.query, because URI.parse
and URI.text above decode and encode the other fields.
SOURCE
/appl/lib/w3c/uris.b
SEE ALSO
charon(1), httpd(8)
Page 3 Plan 9 (printed 10/29/25)