Merge pull request #138 from avsm/uri-3

prepare uri 3

.travis.yml

@@ -7,8 +7,9 @@ script: bash ./.travis-docker.sh
 env:
  global:
    - PACKAGE="uri"
+   - PRE_INSTALL_HOOK="cd /home/opam/opam-repository && git pull origin master && opam update -uy"
    - POST_INSTALL_HOOK="opam --yes depext -yui react ssl lwt"
-   - PINS="uri-sexp:."
+   - PINS="uri.dev:. uri-sexp.dev:."
    - REVDEPS="cohttp git github irmin syndic"
  matrix:
    - DISTRO=centos OCAML_VERSION=4.04

CHANGES.md

@@ -7,9 +7,11 @@ v3.0.0 2019-07-06
   Code that was formerly using `uri.sexp` in its build will now need to
   move to `uri-sexp` instead (#134 @Julow @dinosaure).
 
-* Create a `Uri.Re` module instead of the toplevel `Uri_re`, and
-  deprecate the latter.  The `Uri_re` toplevel module will be removed
-  in the next major release of this library (#134 @Julow @dinosaure).
+* Remove the deprecated `Uri_re` module. All code should be using the
+  `Uri.Re` module instead (@avsm @Julow).
+
+* Remove the `uri.top` library, since we install the toplevel printer
+  automatically since 2.2.0 via an attribute.
 
 v2.2.1 2019-06-02
 -----------------

lib/dune

@@ -1,12 +1,4 @@
 (library
  (name uri)
  (public_name uri)
- (wrapped (transition "Please switch to using Uri.Re instead of Uri_re, which will be removed in Uri.4.0.0 and higher."))
- (modules uri uri_re)
  (libraries re.posix stringext))
-
-(library
- (name uri_top)
- (public_name uri.top)
- (modules uri_top)
- (libraries uri compiler-libs))

lib/uri.ml

@@ -18,6 +18,107 @@
 
 [@@@ocaml.warning "-32"]
 
+module Uri_re = struct
+  open Re
+
+  module Raw = struct
+    let (+) a b = seq [a;b]
+    let (/) a b = alt [a;b]
+
+    let gen_delims = Posix.re "[:/?#\\[\\]@]"
+    let sub_delims = Posix.re "[!$&'()*+,;=]"
+    let c_at = char '@'
+    let c_colon = char ':'
+    let c_slash = char '/'
+    let c_slash2 = Posix.re "//"
+    let c_dot = char '.'
+    let c_question = char '?'
+    let c_hash = char '#'
+
+    let reserved = gen_delims / sub_delims
+    let unreserved = Posix.re "[A-Za-z0-9-._~]"
+    let hexdig = Posix.re "[0-9A-Fa-f]"
+    let pct_encoded = (char '%') + hexdig + hexdig
+
+    let dec_octet = Posix.re "25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?"
+    let ipv4_address = (repn (dec_octet + c_dot) 3 (Some 3)) + dec_octet
+
+    (* following RFC2234, RFC3986, RFC6874 and
+       http://people.spodhuis.org/phil.pennock/software/emit_ipv6_regexp-0.304
+    *)
+    let zone_id = unreserved / pct_encoded
+    let ipv6_address =
+      let (=|) n a = repn a n (Some n) in
+      let (<|) n a = repn a 0 (Some n) in
+      let h16 = repn hexdig 1 (Some 4) in
+      let h16c = h16 + c_colon in
+      let cc = c_colon + c_colon in
+      let ls32 = (h16c + h16) / ipv4_address in
+      ( char '['
+        + (((6=|h16c) + ls32)
+           / (                         cc + (5=|h16c) + ls32)
+           / ((1<|             h16)  + cc + (4=|h16c) + ls32)
+           / ((1<|((1<|h16c) + h16)) + cc + (3=|h16c) + ls32)
+           / ((1<|((2<|h16c) + h16)) + cc + (2=|h16c) + ls32)
+           / ((1<|((3<|h16c) + h16)) + cc +     h16c  + ls32)
+           / ((1<|((4<|h16c) + h16)) + cc             + ls32)
+           / ((1<|((5<|h16c) + h16)) + cc             +  h16)
+           / ((1<|((6<|h16c) + h16)) + cc                   )
+          )
+        + (opt (Posix.re "%25" + rep1 zone_id))
+        + char ']'
+      )
+
+    let reg_name = rep ( unreserved / pct_encoded / sub_delims )
+
+    let host = ipv6_address / ipv4_address / reg_name (* | ipv4_literal TODO *)
+    let userinfo = rep (unreserved / pct_encoded / sub_delims / c_colon)
+    let port = Posix.re "[0-9]*"
+    let authority = (opt ((group userinfo) + c_at)) + (group host) + (opt (c_colon + (group port)))
+    let null_authority = (group empty) + (group empty) + (group empty)
+
+    let pchar = unreserved / pct_encoded / sub_delims / c_colon / c_at
+    let segment = rep pchar
+    let segment_nz = rep1 pchar
+    let segment_nz_nc = repn (unreserved / pct_encoded / sub_delims / c_at) 1 None 
+    let path_abempty = rep (c_slash + segment)
+    let path_absolute = c_slash + (opt (segment_nz + (rep (c_slash + segment))))
+    let path_noscheme = segment_nz_nc + (rep (c_slash + segment ))
+    let path_rootless = segment_nz + (rep (c_slash + segment ))
+    let path_empty = empty
+
+    let path = path_abempty  (* begins with "/" or is empty *)
+               / path_absolute (* begins with "/" but not "//" *)
+               / path_noscheme (* begins with a non-colon segment *)
+               / path_rootless (* begins with a segment *)
+               / path_empty    (* zero characters *)
+
+    let hier_part = (c_slash2 + authority + path_abempty)
+                    / (path_absolute / path_rootless / path_empty)
+
+    let scheme = Posix.re "[A-Za-z][A-Za-z0-9+\\\\-\\.]*"
+    let query = group (rep ( pchar / c_slash / c_question))
+    let fragment = group (rep (pchar / c_slash / c_question))
+
+    let absolute_uri = scheme + c_colon + hier_part + (opt (c_question + query))
+
+    let uri = scheme + c_colon + hier_part + (opt (c_question + query)) + (opt (c_hash + fragment))
+
+    let relative_part = (c_slash2 + authority + path_abempty) / (path_absolute / path_noscheme / path_empty)
+
+    let relative_ref = relative_part + (opt (c_question + query)) + (opt (c_hash + fragment))
+
+    let uri_reference = Posix.re "^(([^:/?#]+):)?(//([^/?#]*))?([^?#]*)(\\?([^#]*))?(#(.*))?"
+  end
+
+  let ipv4_address = Posix.compile Raw.ipv4_address
+  let ipv6_address = Posix.compile Raw.ipv6_address
+  let uri_reference = Posix.compile Raw.uri_reference
+  let authority = Posix.compile Raw.authority
+
+  let host = Posix.compile Raw.host
+end
+
 type component = [
   | `Scheme
   | `Authority
@@ -894,3 +995,6 @@ let canonicalize uri =
 
 let pp ppf uri = Format.pp_print_string ppf (to_string uri)
 let pp_hum ppf uri = Format.pp_print_string ppf (to_string uri)
+
+module Re = Uri_re
+

lib/uri.mli

@@ -258,3 +258,12 @@ val pp : Format.formatter -> t -> unit [@@ocaml.toplevel_printer]
 
 (**  [pp_hum] is now an alias for the {!pp} function. *)
 val pp_hum : Format.formatter -> t -> unit
+
+(** Regular expressions for URI parsing. *)
+module Re : sig
+  val ipv4_address : Re.re
+  val ipv6_address : Re.re
+  val uri_reference : Re.re
+  val authority : Re.re
+  val host : Re.re
+end