mdurl
URL utilities for markdown-it parser.
API
.encode(str [, exclude, keepEncoded]) -> String
Percent-encode a string, avoiding double encoding. Don't touch /a-zA-Z0-9/
+
excluded chars + /%[a-fA-F0-9]{2}/
(if not disabled). Broken surrorates are
replaced with U+FFFD
.
Params:
- str - input string.
-
exclude - optional,
;/?:@&=+$,-_.!~*'()#
. Additional chars to keep intact (except/a-zA-Z0-9/
). -
keepEncoded - optional,
true
. By default it skips already encoded sequences (/%[a-fA-F0-9]{2}/
). If set tofalse
,%
will be encoded.
encode.defaultChars, encode.componentChars
You can use these constants as second argument to encode
function.
-
encode.defaultChars
is the same exclude set as in the standardencodeURI()
function -
encode.componentChars
is the same exclude set as in theencodeURIComponent()
function
For example, encode('something', encode.componentChars, true)
is roughly the equivalent of
the encodeURIComponent()
function (except encode()
doesn't throw).
.decode(str [, exclude]) -> String
Decode percent-encoded string. Invalid percent-encoded sequences (e.g. %2G
)
are left as is. Invalid UTF-8 characters are replaced with U+FFFD
.
Params:
- str - input string.
-
exclude - set of characters to leave encoded, optional,
;/?:@&=+$,#
.
decode.defaultChars, decode.componentChars
You can use these constants as second argument to decode
function.
-
decode.defaultChars
is the same exclude set as in the standarddecodeURI()
function -
decode.componentChars
is the same exclude set as in thedecodeURIComponent()
function
For example, decode('something', decode.defaultChars)
has the same behavior as
decodeURI('something')
on a correctly encoded input.
.parse(url, slashesDenoteHost) -> urlObs
Parse url string. Similar to node's url.parse, but without any normalizations and query string parse.
- url - input url (string)
-
slashesDenoteHost - if url starts with
//
, expect a hostname after it. Optional,false
.
Result (hash):
- protocol
- slashes
- auth
- port
- hostname
- hash
- search
- pathname
Difference with node's url
:
- No leading slash in paths, e.g. in
url.parse('http://foo?bar')
pathname is ``, not/
- Backslashes are not replaced with slashes, so
http:\\example.org\
is treated like a relative path - Trailing colon is treated like a part of the path, i.e. in
http://example.org:foo
pathname is:foo
- Nothing is URL-encoded in the resulting object, (in joyent/node some chars in auth and paths are encoded)
-
url.parse()
does not haveparseQueryString
argument - Removed extraneous result properties:
host
,path
,query
, etc., which can be constructed using other parts of the url.
.format(urlObject)
Format an object previously obtained with .parse()
function. Similar to node's
url.format.