suzanne.soy/racket
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

change -q to config option, fix unlikely race condition in wake-on-signal, add path->url and url->path, and adjust URL parsing

Browse Source 
svn: r9239

original commit: 0153e122b7
This commit is contained in:
Matthew Flatt 2008-04-10 19:05:35 +00:00
parent 4f844dce58
commit 8e59e04b24
2 changed files with 116 additions and 24 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

73
collects/net/scribblings/url.scrbl
View File

@ -73,7 +73,16 @@ URL that ends in a slash. For example, the result of
@scheme[(string->url "http://www.drscheme.org/a/")] has a
@scheme[path] field with strings @scheme["a"] and @scheme[""], while
the result of @scheme[(string->url "http://www.drscheme.org/a")] has a
@scheme[path] field with only the string @scheme["a"].}
@scheme[path] field with only the string @scheme["a"].
When a @scheme["file"] URL is represented by a @scheme[url] structure,
the @scheme[path] field is mostly a list of path elements. For Unix
paths, the root directory is not included in @scheme[path]; its
presence or absence is implicit in the @scheme[path-absolute?] flag.
For Windows paths, the first element typically represents a drive, but
a UNC path is represented by a first element that is @scheme[""] and
then successive elements complete the drive components that are
separated by @litchar{/} or @litchar{\}.}
@defstruct[path/param ([path (or/c string? (one-of/c 'up 'same))]
[param (listof string?)])]{
@ -95,7 +104,26 @@ Parses the URL specified by @scheme[str] into a @scheme[url]
struct. The @scheme[string->url] procedure uses
@scheme[form-urlencoded->alist] when parsing the query, so it is
sensitive to the @scheme[current-alist-separator-mode] parameter for
determining the association separator.}
determining the association separator.
If @scheme[str] starts with @scheme["file:"], then the path is always
parsed as an absolute path, and the parsing details depend on
@scheme[file-url-path-convention-type]:
@itemize[
@item{@scheme['unix] : If @scheme["file:"] is followed by
@litchar{//} and a non-@litchar{/}, then the first element
after the @litchar{//} is parsed as a host (and maybe port);
otherwise, the first element starts the path, and the host is
@scheme[""].}
@item{@scheme['windows] : If @scheme["file:"] is followed by
@litchar{//}, then the @litchar{//} is stripped; the remainder
parsed as a Windows path. The host is always @scheme[""] and
the port is always @scheme[#f].}
]}
@defproc[(combine-url/relative [base url?] [relative string?]) url?]{
@ -117,17 +145,50 @@ scheme @scheme["http"].}
@defproc[(url->string [URL url?]) string?]{
Generates a string corresponding to the contents of a @scheme[url] struct.
For a @scheme["file:"] URL, empty strings in the path list are treated as
@scheme['same] for @scheme[build-path].
Generates a string corresponding to the contents of a @scheme[url]
struct. For a @scheme["file:"] URL, the URL must not be relative, the
result always starts @litchar{file://}, and the interpretation of the
path depends on the value of @scheme[file-url-path-convention-type]:
@itemize[
@item{@scheme['unix] : Elements in @scheme[URL] are treated as path
elements. Empty strings in the path list are treated like
@scheme['same].}
@item{@scheme['windows] : If the first element is @scheme[""] then
the next two elements define the UNC root, and the rest of the
elements are treated as path elements. Empty strings in the
path list are treated like @scheme['same].}
]
The @scheme[url->string] procedure uses
@scheme[alist->form-urlencoded] when formatting the query, so it it
@scheme[alist->form-urlencoded] when formatting the query, so it is
sensitive to the @scheme[current-alist-separator-mode] parameter for
determining the association separator. The default is to separate
associations with a @litchar{&}.}
@defproc[(path->url [path (or/c path-string? path-for-some-system?)])
url?]{
Converts a path to a @scheme[url].}
@defproc[(url->path [URL url?]
[kind (one-of/c 'unix 'windows) (system-path-convention-type)])
path-for-some-system?]{
Converts @scheme[URL], which is assumed to be a @scheme["file"] URL,
to a path.}
@defparam[file-url-path-convention-type kind (one-of/c 'unix 'windows)]{
Determines the default conversion to and from strings for
@scheme["file"] URLs. See @scheme[string->url] and @scheme[url->string].}
@deftogether[(
@defproc[(get-pure-port [URL url?]

67
collects/tests/mzscheme/net.ss
View File

@ -11,7 +11,6 @@
(require net/url
net/uri-codec
mzlib/string
net/url-unit ; to get set-url:os-type!
)
(test "%Pq" uri-decode "%Pq")
@ -280,8 +279,10 @@
(err/rt-test (string->url "a b://www.foo.com/") url-exception?)
;; test file: urls
(test-s->u #("file" #f #f #f #t (#("abc") #("def.html")) () #f)
"file:/abc/def.html")
(test-s->u #("file" #f "" #f #t (#("abc") #("def.html")) () #f)
"file:///abc/def.html")
(test "file:///abc/def.html" url->string (string->url "file:///abc/def.html"))
(test "file://a/b" url->string (string->url "file://a/b"))
(test-s->u #("file" #f "localhost" #f #t (#("abc") #("def.html")) () #f)
"file://localhost/abc/def.html")
@ -289,26 +290,56 @@
;; test files: urls with colons, and the different parsing on Windows
(test-s->u #("file" #f "localhost" 123 #t (#("abc") #("def.html")) () #f)
"file://localhost:123/abc/def.html")
(set-url:os-type! 'unix)
;; different parse for file://foo:/...
(test #("file" #f "foo" #f #t (#("abc") #("def.html")) () #f)
(parameterize ([file-url-path-convention-type 'unix])
;; different parse for file://foo:/...
(test #("file" #f "foo" #f #t (#("abc") #("def.html")) () #f)
string->url/vec
"file://foo:/abc/def.html"))
(parameterize ([file-url-path-convention-type 'windows])
(test #("file" #f "" #f #t (#("foo:") #("abc") #("def.html")) () #f)
string->url/vec
"file://foo:/abc/def.html")
(set-url:os-type! 'windows)
(test #("file" #f #f #f #f (#("foo:") #("abc") #("def.html")) () #f)
(test #("file" #f "" #f #t (#("c:") #("abc") #("def.html")) () #f)
string->url/vec
"file://foo:/abc/def.html")
(set-url:os-type! 'unix)
;; but no effect on http://foo:/...
(test #("http" #f "foo" #f #t (#("abc") #("def.html")) () #f)
"file://c:/abc/def.html")
(test #("file" #f "" #f #t (#("") #("d") #("c") #("abc") #("def.html")) () #f)
string->url/vec
"http://foo:/abc/def.html")
(set-url:os-type! 'windows)
(test #("http" #f "foo" #f #t (#("abc") #("def.html")) () #f)
string->url/vec
"http://foo:/abc/def.html")
(set-url:os-type! 'unix)
"file:\\\\d\\c\\abc\\def.html"))
(parameterize ([file-url-path-convention-type 'unix])
;; but no effect on http://foo:/...
(test #("http" #f "foo" #f #t (#("abc") #("def.html")) () #f)
string->url/vec
"http://foo:/abc/def.html"))
(parameterize ([file-url-path-convention-type 'windows])
(test #("http" #f "foo" #f #t (#("abc") #("def.html")) () #f)
string->url/vec
"http://foo:/abc/def.html"))
(test "file:///c:/a/b"
url->string (path->url (bytes->path #"c:\\a\\b" 'windows)))
(test "file:///c:/a/b"
url->string (path->url (bytes->path #"\\\\?\\c:\\a\\b" 'windows)))
(test #"/a/b/c" path->bytes
(url->path (path->url (bytes->path #"/a/b/c" 'unix)) 'unix))
(test #"a/b/c" path->bytes
(url->path (path->url (bytes->path #"a/b/c" 'unix)) 'unix))
(test #"c:\\a\\b" path->bytes
(url->path (path->url (bytes->path #"c:/a/b" 'windows)) 'windows))
(test #"a\\b" path->bytes
(url->path (path->url (bytes->path #"a/b" 'windows)) 'windows))
(test #"\\\\d\\c\\a" path->bytes
(url->path (path->url (bytes->path #"//d/c/a" 'windows)) 'windows))
(test #"c:\\a\\b" path->bytes
(url->path (path->url (bytes->path #"\\\\?\\c:\\a\\b" 'windows)) 'windows))
(test #"\\\\d\\c\\a\\b" path->bytes
(url->path (path->url (bytes->path #"\\\\?\\UNC\\d\\c\\a\\b" 'windows)) 'windows))
(test #"\\\\?\\c:\\a/x\\b" path->bytes
(url->path (path->url (bytes->path #"\\\\?\\c:\\a/x\\b" 'windows)) 'windows))
(test #"\\\\?\\UNC\\d\\c\\a/x\\b" path->bytes
(url->path (path->url (bytes->path #"\\\\?\\UNC\\d\\\\c\\a/x\\b" 'windows)) 'windows))
;; see PR8809 (value-less keys in the query part)
(test-s->u #("http" #f "foo.bar" #f #t (#("baz")) ((ugh . #f)) #f)
"http://foo.bar/baz?ugh")