refactor: Moved the order of functions around to follow the ABNF doc

This commit is contained in:
2025-09-14 17:51:31 +01:00
parent 8b8d3e577e
commit 47da8071cf
3 changed files with 513 additions and 443 deletions

View File

@@ -1,13 +1,18 @@
import gleam/bool
import gleam/int import gleam/int
import gleam/list.{Continue, Stop} import gleam/list.{Continue, Stop}
import gleam/option.{None, Some} import gleam/option.{None, Some}
import gleam/result import gleam/result
import gleam/string import gleam/string
import gleam/uri.{type Uri, Uri, empty} import gleam/uri.{type Uri, Uri, empty}
import gluri/internal/utils import gluri/internal/utils.{
combine_uris, parse_hex_digit, parse_hex_digits, parse_min_max, parse_optional,
parse_this_then,
}
import splitter import splitter
// URI-reference = URI / relative-ref
// URI = scheme ":" hier-part [ "?" query ] [ "#" fragment ]
// relative-ref = relative-part [ "?" query ] [ "#" fragment ]
pub fn parse(uri: String) -> Result(Uri, Nil) { pub fn parse(uri: String) -> Result(Uri, Nil) {
case parse_scheme(uri) { case parse_scheme(uri) {
Ok(#(scheme, rest)) -> { Ok(#(scheme, rest)) -> {
@@ -37,6 +42,23 @@ pub fn parse(uri: String) -> Result(Uri, Nil) {
} }
} }
// hier-part = "//" authority path-abempty
// / path-absolute
// / path-rootless
// / path-empty
fn parse_hier_part(str: String) -> Result(#(Uri, String), Nil) {
utils.try_parsers(
[
parse_authority,
parse_path_absolute,
parse_path_rootless,
parse_path_empty,
],
str,
)
}
// query = *( pchar / "/" / "?" )
fn parse_query(str: String) -> Result(#(Uri, String), Nil) { fn parse_query(str: String) -> Result(#(Uri, String), Nil) {
case str { case str {
"?" <> rest -> { "?" <> rest -> {
@@ -48,6 +70,7 @@ fn parse_query(str: String) -> Result(#(Uri, String), Nil) {
} }
} }
// fragment = *( pchar / "/" / "?" )
fn parse_fragment(str: String) -> Result(#(Uri, String), Nil) { fn parse_fragment(str: String) -> Result(#(Uri, String), Nil) {
case str { case str {
"#" <> rest -> { "#" <> rest -> {
@@ -59,98 +82,80 @@ fn parse_fragment(str: String) -> Result(#(Uri, String), Nil) {
} }
} }
fn parse_hier_part(str: String) -> Result(#(Uri, String), Nil) { // relative-part = "//" authority path-abempty
utils.try_parsers( // / path-absolute
[parse_authority, parse_absolute, parse_rootless, parse_empty], // / path-noscheme
str, // / path-empty
)
}
fn parse_relative_part(str: String) -> Result(#(Uri, String), Nil) { fn parse_relative_part(str: String) -> Result(#(Uri, String), Nil) {
utils.try_parsers( utils.try_parsers(
[parse_authority, parse_absolute, parse_noscheme, parse_empty], [
parse_authority,
parse_path_absolute,
parse_path_noscheme,
parse_path_empty,
],
str, str,
) )
} }
fn parse_absolute(str: String) -> Result(#(Uri, String), Nil) { // scheme = ALPHA *( ALPHA / DIGIT / "+" / "-" / "." )
fn parse_scheme(str: String) -> Result(#(Uri, String), Nil) {
case str { case str {
"/" <> rest -> { "http:" <> rest ->
use #(seg, rest) <- result.try( Ok(#(Uri(Some("http"), None, None, None, "", None, None), rest))
parse_optional(rest, parse_this_then( "https:" <> rest ->
Ok(#(Uri(Some("https"), None, None, None, "", None, None), rest))
"ftp:" <> rest ->
Ok(#(Uri(Some("ftp"), None, None, None, "", None, None), rest))
"file:" <> rest ->
Ok(#(Uri(Some("file"), None, None, None, "", None, None), rest))
"ws:" <> rest ->
Ok(#(Uri(Some("ws"), None, None, None, "", None, None), rest))
"wss:" <> rest ->
Ok(#(Uri(Some("wss"), None, None, None, "", None, None), rest))
_ -> {
case parse_alpha(str) {
Ok(#(first, rest)) -> {
case do_parse_scheme(rest, first) {
Error(_) -> Error(Nil)
Ok(#(scheme, rest)) ->
Ok(#(Uri(Some(scheme), None, None, None, "", None, None), rest))
}
}
_ -> Error(Nil)
}
}
}
}
fn do_parse_scheme(
str: String,
scheme: String,
) -> Result(#(String, String), Nil) {
case str {
":" <> rest -> Ok(#(scheme, rest))
"" -> Error(Nil)
_ -> {
use #(part, rest) <- result.try(utils.try_parsers(
[ [
do_parse_segment_nz, parse_alpha,
utils.get_multiple_optional_result( parse_digit,
fn(str) { fn(str) {
case str { case str {
"/" <> rest -> { "+" as l <> rest | "-" as l <> rest | "." as l <> rest ->
do_parse_segment(rest, do_parse_pchar, "/") Ok(#(l, rest))
}
_ -> Error(Nil) _ -> Error(Nil)
} }
}, },
_,
),
], ],
_, str,
)), ))
) do_parse_scheme(rest, scheme <> part)
Ok(#(Uri(None, None, None, None, "/" <> seg, None, None), rest))
} }
_ -> Error(Nil)
} }
} }
fn parse_rootless(str: String) -> Result(#(Uri, String), Nil) { // authority = [ userinfo "@" ] host [ ":" port ]
use #(seg1, rest) <- result.try(do_parse_segment_nz(str))
let #(segs, rest) =
utils.get_multiple_optional(
fn(str) {
case str {
"/" <> rest -> {
do_parse_segment(rest, do_parse_pchar, "/")
}
_ -> Error(Nil)
}
},
rest,
)
Ok(#(Uri(None, None, None, None, seg1 <> segs, None, None), rest))
}
fn parse_noscheme(str: String) -> Result(#(Uri, String), Nil) {
use #(seg1, rest) <- result.try(do_parse_segment_nz_nc(str))
let #(segs, rest) =
utils.get_multiple_optional(
fn(str) {
case str {
"/" <> rest -> {
do_parse_segment(rest, do_parse_pchar, "/")
}
_ -> Error(Nil)
}
},
rest,
)
Ok(#(Uri(None, None, None, None, seg1 <> segs, None, None), rest))
}
fn parse_optional(str, opt_fn) {
case opt_fn(str) {
Error(Nil) -> Ok(#("", str))
Ok(r) -> Ok(r)
}
}
fn parse_empty(str: String) -> Result(#(Uri, String), Nil) {
Ok(#(Uri(None, None, None, None, "", None, None), str))
}
fn parse_authority(str: String) -> Result(#(Uri, String), Nil) { fn parse_authority(str: String) -> Result(#(Uri, String), Nil) {
case str { case str {
"//" <> rest -> { "//" <> rest -> {
@@ -181,6 +186,40 @@ fn parse_authority_part(str: String) -> Result(#(Uri, String), Nil) {
Ok(#(Uri(None, ui, Some(host), port, path, None, None), rest)) Ok(#(Uri(None, ui, Some(host), port, path, None, None), rest))
} }
// userinfo = *( unreserved / pct-encoded / sub-delims / ":" )
fn parse_userinfo(
str: String,
userinfo: String,
) -> Result(#(String, String), Nil) {
case str {
"@" <> rest -> Ok(#(userinfo, rest))
"" -> Error(Nil)
_ -> {
use #(part, rest) <- result.try(utils.try_parsers(
[
parse_unreserved,
parse_pct_encoded,
parse_sub_delim,
fn(str: String) {
case str {
":" as l <> rest -> Ok(#(l, rest))
_ -> Error(Nil)
}
},
],
str,
))
parse_userinfo(rest, userinfo <> part)
}
}
}
// host = IP-literal / IPv4address / reg-name
fn parse_host(str: String) {
utils.try_parsers([parse_ip_literal, parse_ipv4address, parse_reg_name], str)
}
// port = *DIGIT
fn parse_port(str: String) { fn parse_port(str: String) {
case str { case str {
":" <> rest -> { ":" <> rest -> {
@@ -190,24 +229,12 @@ fn parse_port(str: String) {
} }
} }
fn parse_digits(str: String, digits: String) { // IP-literal = "[" ( IPv6address / IPvFuture ) "]"
case parse_digit(str) {
Ok(#(d, rest)) -> {
parse_digits(rest, digits <> d)
}
Error(_) -> #(digits, str)
}
}
fn parse_host(str: String) {
utils.try_parsers([parse_ip_literal, parse_ipv4, parse_reg_name], str)
}
fn parse_ip_literal(str: String) { fn parse_ip_literal(str: String) {
case str { case str {
"[" <> rest -> { "[" <> rest -> {
use #(ip, rest) <- result.try(utils.try_parsers( use #(ip, rest) <- result.try(utils.try_parsers(
[parse_ipv6, parse_ipfuture], [parse_ipv6, parse_ipvfuture],
rest, rest,
)) ))
case rest { case rest {
@@ -219,6 +246,50 @@ fn parse_ip_literal(str: String) {
} }
} }
// IPvFuture = "v" 1*HEXDIG "." 1*( unreserved / sub-delims / ":" )
fn parse_ipvfuture(str: String) {
case str {
"v" <> rest -> {
use #(v, rest) <- result.try(utils.get_multiple(parse_hex_digit, rest))
case rest {
"." <> rest -> {
use #(i, rest) <- result.try(utils.get_multiple(
fn(str) {
utils.try_parsers(
[
parse_unreserved,
parse_sub_delim,
fn(str: String) {
case str {
":" as l <> rest -> Ok(#(l, rest))
_ -> Error(Nil)
}
},
],
str,
)
},
rest,
))
Ok(#("v" <> v <> "." <> i, rest))
}
_ -> Error(Nil)
}
}
_ -> Error(Nil)
}
}
// IPv6address = 6( h16 ":" ) ls32
// / "::" 5( h16 ":" ) ls32
// / [ h16 ] "::" 4( h16 ":" ) ls32
// / [ *1( h16 ":" ) h16 ] "::" 3( h16 ":" ) ls32
// / [ *2( h16 ":" ) h16 ] "::" 2( h16 ":" ) ls32
// / [ *3( h16 ":" ) h16 ] "::" h16 ":" ls32
// / [ *4( h16 ":" ) h16 ] "::" ls32
// / [ *5( h16 ":" ) h16 ] "::" h16
// / [ *6( h16 ":" ) h16 ] "::"
fn parse_ipv6(str: String) { fn parse_ipv6(str: String) {
utils.try_parsers( utils.try_parsers(
[ [
@@ -291,10 +362,6 @@ fn parse_ipv6(str: String) {
) )
} }
fn parse_h16s(str: String, max) {
parse_min_max(str, 0, max, parse_h16_colon)
}
fn parse_colons(str: String) { fn parse_colons(str: String) {
case str { case str {
"::" <> rest -> Ok(#("::", rest)) "::" <> rest -> Ok(#("::", rest))
@@ -302,23 +369,13 @@ fn parse_colons(str: String) {
} }
} }
fn parse_this_then( // h16 = 1*4HEXDIG
parsers: List(fn(String) -> Result(#(String, String), Nil)), fn parse_h16(str: String) {
str: String, parse_hex_digits(str, 1, 4)
) {
list.fold_until(parsers, Ok(#("", str)), fn(acc, parser) {
let assert Ok(#(res, str)) = acc
case parser(str) {
Ok(#(res2, rest)) -> {
Continue(Ok(#(res <> res2, rest)))
}
Error(Nil) -> Stop(Error(Nil))
}
})
} }
fn parse_ls32(str: String) -> Result(#(String, String), Nil) { fn parse_h16s(str: String, max) {
utils.try_parsers([parse_h16_pair, parse_ipv4], str) parse_min_max(str, 0, max, parse_h16_colon)
} }
fn parse_h16_pair(str: String) { fn parse_h16_pair(str: String) {
@@ -332,10 +389,6 @@ fn parse_h16_pair(str: String) {
} }
} }
fn parse_h16(str: String) {
parse_hex_digits(str, 1, 4)
}
fn parse_h16_colon(str: String) { fn parse_h16_colon(str: String) {
use #(h16, rest) <- result.try(parse_h16(str)) use #(h16, rest) <- result.try(parse_h16(str))
case rest { case rest {
@@ -344,188 +397,13 @@ fn parse_h16_colon(str: String) {
} }
} }
fn parse_ipfuture(str: String) { // ls32 = ( h16 ":" h16 ) / IPv4address
case str { fn parse_ls32(str: String) -> Result(#(String, String), Nil) {
"v" <> rest -> { utils.try_parsers([parse_h16_pair, parse_ipv4address], str)
use #(v, rest) <- result.try(utils.get_multiple(
utils.parse_hex_digit,
rest,
))
case rest {
"." <> rest -> {
use #(i, rest) <- result.try(utils.get_multiple(
fn(str) {
utils.try_parsers(
[
parse_unreserved,
parse_sub_delim,
fn(str: String) {
case str {
":" as l <> rest -> Ok(#(l, rest))
_ -> Error(Nil)
}
},
],
str,
)
},
rest,
))
Ok(#("v" <> v <> "." <> i, rest))
}
_ -> Error(Nil)
}
}
_ -> Error(Nil)
}
} }
fn parse_query_fragment(str: String) { // IPv4address = dec-octet "." dec-octet "." dec-octet "." dec-octet
utils.try_parsers( fn parse_ipv4address(str: String) {
[
do_parse_pchar,
fn(str: String) {
case str {
"/" as l <> rest | "?" as l <> rest -> Ok(#(l, rest))
_ -> Error(Nil)
}
},
],
str,
)
}
fn parse_path_abempty(str: String) -> #(String, String) {
utils.get_multiple_optional(
fn(str) {
case str {
"/" <> rest -> {
do_parse_segment(rest, do_parse_pchar, "/")
}
_ -> Error(Nil)
}
},
str,
)
}
fn do_parse_segment(
str: String,
char_fn,
segment: String,
) -> Result(#(String, String), Nil) {
case char_fn(str) {
Error(Nil) | Ok(#("", _)) -> Ok(#(segment, str))
Ok(#(l, rest)) -> do_parse_segment(rest, char_fn, segment <> l)
}
}
fn do_parse_segment_nz(str: String) {
use #(char1, rest) <- result.try(do_parse_pchar(str))
use #(chars, rest) <- result.try(do_parse_segment(rest, do_parse_pchar, char1))
Ok(#(chars, rest))
}
fn do_parse_segment_nz_nc(str: String) {
use #(char1, rest) <- result.try(do_parse_pchar_nc(str))
use #(chars, rest) <- result.try(do_parse_segment(
rest,
do_parse_pchar_nc,
char1,
))
Ok(#(chars, rest))
}
fn do_parse_pchar(str: String) {
utils.try_parsers(
[
parse_unreserved,
parse_pct_encoded,
parse_sub_delim,
fn(str: String) {
case str {
":" as l <> rest | "@" as l <> rest -> Ok(#(l, rest))
_ -> Error(Nil)
}
},
],
str,
)
}
fn do_parse_pchar_nc(str: String) {
utils.try_parsers(
[
parse_unreserved,
parse_pct_encoded,
parse_sub_delim,
fn(str: String) {
case str {
"@" as l <> rest -> Ok(#(l, rest))
_ -> Error(Nil)
}
},
],
str,
)
}
pub fn parse_reg_name(str: String) {
// can't error
case do_parse_reg_name(str, "") {
Error(Nil) -> Ok(#("", str))
Ok(#(reg_name, rest)) -> Ok(#(reg_name, rest))
}
}
fn do_parse_reg_name(str: String, reg_name: String) {
case
utils.try_parsers(
[parse_unreserved, parse_pct_encoded, parse_sub_delim],
str,
)
{
Error(Nil) | Ok(#("", _)) -> Ok(#(reg_name, str))
Ok(#(l, rest)) -> do_parse_reg_name(rest, reg_name <> l)
}
}
fn parse_pct_encoded(str: String) {
case str {
"%" <> rest -> {
use #(hex1, rest) <- result.try(utils.parse_hex_digit(rest))
use #(hex2, rest) <- result.try(utils.parse_hex_digit(rest))
Ok(#("%" <> hex1 <> hex2, rest))
}
_ -> Error(Nil)
}
}
fn parse_sub_delim(str: String) {
case str {
"!" as l <> rest
| "$" as l <> rest
| "&" as l <> rest
| "'" as l <> rest
| "(" as l <> rest
| ")" as l <> rest
| "*" as l <> rest
| "+" as l <> rest
| "," as l <> rest
| ";" as l <> rest
| "=" as l <> rest -> Ok(#(l, rest))
_ -> Error(Nil)
}
}
fn parse_ipv4(str: String) {
use #(oct1, rest) <- result.try(parse_dec_octet(str)) use #(oct1, rest) <- result.try(parse_dec_octet(str))
use rest <- result.try(case rest { use rest <- result.try(case rest {
"." <> rest -> Ok(rest) "." <> rest -> Ok(rest)
@@ -553,6 +431,11 @@ const octet_matches = [
["0123456789"], ["0123456789"],
] ]
// dec-octet = DIGIT ; 0-9
// / %x31-39 DIGIT ; 10-99
// / "1" 2DIGIT ; 100-199
// / "2" %x30-34 DIGIT ; 200-249
// / "25" %x30-35 ; 250-255
fn parse_dec_octet(str: String) -> Result(#(String, String), Nil) { fn parse_dec_octet(str: String) -> Result(#(String, String), Nil) {
list.fold_until(octet_matches, Error(Nil), fn(_, chars) { list.fold_until(octet_matches, Error(Nil), fn(_, chars) {
case case
@@ -575,114 +458,265 @@ fn parse_dec_octet(str: String) -> Result(#(String, String), Nil) {
}) })
} }
fn parse_userinfo( // reg-name = *( unreserved / pct-encoded / sub-delims )
str: String, pub fn parse_reg_name(str: String) {
userinfo: String, // can't error
) -> Result(#(String, String), Nil) {
case do_parse_reg_name(str, "") {
Error(Nil) -> Ok(#("", str))
Ok(#(reg_name, rest)) -> Ok(#(reg_name, rest))
}
}
fn do_parse_reg_name(str: String, reg_name: String) {
case
utils.try_parsers(
[parse_unreserved, parse_pct_encoded, parse_sub_delim],
str,
)
{
Error(Nil) | Ok(#("", _)) -> Ok(#(reg_name, str))
Ok(#(l, rest)) -> do_parse_reg_name(rest, reg_name <> l)
}
}
// 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
// path-abempty = *( "/" segment )
fn parse_path_abempty(str: String) -> #(String, String) {
utils.get_multiple_optional(
fn(str) {
case str { case str {
"@" <> rest -> Ok(#(userinfo, rest)) "/" <> rest -> {
"" -> Error(Nil) do_parse_segment(rest, do_parse_pchar, "/")
_ -> { }
use #(part, rest) <- result.try(utils.try_parsers( _ -> Error(Nil)
}
},
str,
)
}
// path-absolute = "/" [ segment-nz *( "/" segment ) ]
fn parse_path_absolute(str: String) -> Result(#(Uri, String), Nil) {
case str {
"/" <> rest -> {
use #(seg, rest) <- result.try(
parse_optional(rest, parse_this_then(
[
do_parse_segment_nz,
utils.get_multiple_optional_result(
fn(str) {
case str {
"/" <> rest -> {
do_parse_segment(rest, do_parse_pchar, "/")
}
_ -> Error(Nil)
}
},
_,
),
],
_,
)),
)
Ok(#(Uri(None, None, None, None, "/" <> seg, None, None), rest))
}
_ -> Error(Nil)
}
}
// path-noscheme = segment-nz-nc *( "/" segment )
fn parse_path_noscheme(str: String) -> Result(#(Uri, String), Nil) {
use #(seg1, rest) <- result.try(do_parse_segment_nz_nc(str))
let #(segs, rest) =
utils.get_multiple_optional(
fn(str) {
case str {
"/" <> rest -> {
do_parse_segment(rest, do_parse_pchar, "/")
}
_ -> Error(Nil)
}
},
rest,
)
Ok(#(Uri(None, None, None, None, seg1 <> segs, None, None), rest))
}
// path-rootless = segment-nz *( "/" segment )
fn parse_path_rootless(str: String) -> Result(#(Uri, String), Nil) {
use #(seg1, rest) <- result.try(do_parse_segment_nz(str))
let #(segs, rest) =
utils.get_multiple_optional(
fn(str) {
case str {
"/" <> rest -> {
do_parse_segment(rest, do_parse_pchar, "/")
}
_ -> Error(Nil)
}
},
rest,
)
Ok(#(Uri(None, None, None, None, seg1 <> segs, None, None), rest))
}
// path-empty = 0<pchar>
fn parse_path_empty(str: String) -> Result(#(Uri, String), Nil) {
Ok(#(Uri(None, None, None, None, "", None, None), str))
}
// segment = *pchar
fn do_parse_segment(
str: String,
char_fn,
segment: String,
) -> Result(#(String, String), Nil) {
case char_fn(str) {
Error(Nil) | Ok(#("", _)) -> Ok(#(segment, str))
Ok(#(l, rest)) -> do_parse_segment(rest, char_fn, segment <> l)
}
}
// segment-nz = 1*pchar
fn do_parse_segment_nz(str: String) {
use #(char1, rest) <- result.try(do_parse_pchar(str))
use #(chars, rest) <- result.try(do_parse_segment(rest, do_parse_pchar, char1))
Ok(#(chars, rest))
}
// segment-nz-nc = 1*( unreserved / pct-encoded / sub-delims / "@" )
// ; non-zero-length segment without any colon ":"
fn do_parse_segment_nz_nc(str: String) {
use #(char1, rest) <- result.try(do_parse_pchar_without_colon(str))
use #(chars, rest) <- result.try(do_parse_segment(
rest,
do_parse_pchar_without_colon,
char1,
))
Ok(#(chars, rest))
}
// pchar = unreserved / pct-encoded / sub-delims / ":" / "@"
fn do_parse_pchar(str: String) {
utils.try_parsers(
[ [
parse_unreserved, parse_unreserved,
parse_pct_encoded, parse_pct_encoded,
parse_sub_delim, parse_sub_delim,
fn(str: String) { fn(str: String) {
case str { case str {
":" as l <> rest -> Ok(#(l, rest)) ":" as l <> rest | "@" as l <> rest -> Ok(#(l, rest))
_ -> Error(Nil) _ -> Error(Nil)
} }
}, },
], ],
str, str,
)) )
parse_userinfo(rest, userinfo <> part)
}
}
} }
fn parse_scheme(str: String) -> Result(#(Uri, String), Nil) { fn do_parse_pchar_without_colon(str: String) {
utils.try_parsers(
[
parse_unreserved,
parse_pct_encoded,
parse_sub_delim,
fn(str: String) {
case str { case str {
"http:" <> rest -> "@" as l <> rest -> Ok(#(l, rest))
Ok(#(Uri(Some("http"), None, None, None, "", None, None), rest)) _ -> Error(Nil)
"https:" <> rest ->
Ok(#(Uri(Some("https"), None, None, None, "", None, None), rest))
"ftp:" <> rest ->
Ok(#(Uri(Some("ftp"), None, None, None, "", None, None), rest))
"file:" <> rest ->
Ok(#(Uri(Some("file"), None, None, None, "", None, None), rest))
"ws:" <> rest ->
Ok(#(Uri(Some("ws"), None, None, None, "", None, None), rest))
"wss:" <> rest ->
Ok(#(Uri(Some("wss"), None, None, None, "", None, None), rest))
_ -> {
case parse_alpha(str) {
Ok(#(first, rest)) -> {
case do_parse_scheme(rest, first) {
Error(_) -> Error(Nil)
Ok(#(scheme, rest)) ->
Ok(#(Uri(Some(scheme), None, None, None, "", None, None), rest))
} }
},
],
str,
)
}
// query = *( pchar / "/" / "?" )
// fragment = *( pchar / "/" / "?" )
fn parse_query_fragment(str: String) {
utils.try_parsers(
[
do_parse_pchar,
fn(str: String) {
case str {
"/" as l <> rest | "?" as l <> rest -> Ok(#(l, rest))
_ -> Error(Nil)
}
},
],
str,
)
}
// pct-encoded = "%" HEXDIG HEXDIG
fn parse_pct_encoded(str: String) {
case str {
"%" <> rest -> {
use #(hex1, rest) <- result.try(parse_hex_digit(rest))
use #(hex2, rest) <- result.try(parse_hex_digit(rest))
Ok(#("%" <> hex1 <> hex2, rest))
} }
_ -> Error(Nil) _ -> Error(Nil)
} }
} }
}
}
fn do_parse_scheme( // unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~"
str: String, // reserved = gen-delims / sub-delims
scheme: String, // gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@"
) -> Result(#(String, String), Nil) { fn parse_unreserved(str: String) -> Result(#(String, String), Nil) {
case str { utils.try_parsers(
":" <> rest -> Ok(#(scheme, rest))
"" -> Error(Nil)
_ -> {
use #(part, rest) <- result.try(utils.try_parsers(
[ [
parse_alpha, parse_alpha,
parse_digit, parse_digit,
fn(str) { fn(str) {
case str { case str {
"+" as l <> rest | "-" as l <> rest | "." as l <> rest -> "_" as l <> rest
Ok(#(l, rest)) | "-" as l <> rest
| "." as l <> rest
| "~" as l <> rest -> Ok(#(l, rest))
_ -> Error(Nil) _ -> Error(Nil)
} }
}, },
], ],
str, str,
)) )
do_parse_scheme(rest, scheme <> part)
} }
// sub-delims = "!" / "$" / "&" / "'" / "(" / ")"
// / "*" / "+" / "," / ";" / "="
fn parse_sub_delim(str: String) {
case str {
"!" as l <> rest
| "$" as l <> rest
| "&" as l <> rest
| "'" as l <> rest
| "(" as l <> rest
| ")" as l <> rest
| "*" as l <> rest
| "+" as l <> rest
| "," as l <> rest
| ";" as l <> rest
| "=" as l <> rest -> Ok(#(l, rest))
_ -> Error(Nil)
} }
} }
fn parse_min_max(str, min, max, parse_fn) { // DIGIT = %x3039
use <- bool.guard(when: min < 0 || max <= 0 || min > max, return: Error(Nil))
case
list.repeat("", max)
|> list.fold_until(Ok(#("", str, 0)), fn(acc, _) {
let assert Ok(#(hex, str, i)) = acc
case parse_fn(str) {
Error(_) ->
case i < min {
True -> Stop(Error(Nil))
False -> Stop(Ok(#(hex, str, i)))
}
Ok(#(l, rest)) -> Continue(Ok(#(hex <> l, rest, i + 1)))
}
})
{
Error(_) -> Error(Nil)
Ok(#(hex, str, _)) -> Ok(#(hex, str))
}
}
fn parse_hex_digits(str, min, max) {
parse_min_max(str, min, max, utils.parse_hex_digit)
}
fn parse_digit(str: String) -> Result(#(String, String), Nil) { fn parse_digit(str: String) -> Result(#(String, String), Nil) {
case str { case str {
"0" as l <> rest "0" as l <> rest
@@ -699,6 +733,16 @@ fn parse_digit(str: String) -> Result(#(String, String), Nil) {
} }
} }
fn parse_digits(str: String, digits: String) {
case parse_digit(str) {
Ok(#(d, rest)) -> {
parse_digits(rest, digits <> d)
}
Error(_) -> #(digits, str)
}
}
// ALPHA = %x415A | %x617A
fn parse_alpha(str: String) -> Result(#(String, String), Nil) { fn parse_alpha(str: String) -> Result(#(String, String), Nil) {
case str { case str {
"a" as l <> rest "a" as l <> rest
@@ -757,60 +801,6 @@ fn parse_alpha(str: String) -> Result(#(String, String), Nil) {
} }
} }
fn parse_unreserved(str: String) -> Result(#(String, String), Nil) {
utils.try_parsers(
[
parse_alpha,
parse_digit,
fn(str) {
case str {
"_" as l <> rest
| "-" as l <> rest
| "." as l <> rest
| "~" as l <> rest -> Ok(#(l, rest))
_ -> Error(Nil)
}
},
],
str,
)
}
fn combine_uris(uris: List(Uri)) -> Uri {
list.fold(uris, Uri(None, None, None, None, "", None, None), fn(acc, uri) {
let acc = case uri {
Uri(Some(scheme), _, _, _, _, _, _) -> Uri(..acc, scheme: Some(scheme))
_ -> acc
}
let acc = case uri {
Uri(_, Some(userinfo), _, _, _, _, _) ->
Uri(..acc, userinfo: Some(userinfo))
_ -> acc
}
let acc = case uri {
Uri(_, _, Some(host), _, _, _, _) -> Uri(..acc, host: Some(host))
_ -> acc
}
let acc = case uri {
Uri(_, _, _, Some(port), _, _, _) -> Uri(..acc, port: Some(port))
_ -> acc
}
let acc = case uri {
Uri(_, _, _, _, path, _, _) if path != "" -> Uri(..acc, path: path)
_ -> acc
}
let acc = case uri {
Uri(_, _, _, _, _, Some(query), _) -> Uri(..acc, query: Some(query))
_ -> acc
}
case uri {
Uri(_, _, _, _, _, _, Some(fragment)) ->
Uri(..acc, fragment: Some(fragment))
_ -> acc
}
})
}
pub fn parse_query_parts(query: String) -> Result(List(#(String, String)), Nil) { pub fn parse_query_parts(query: String) -> Result(List(#(String, String)), Nil) {
let splitter = splitter.new(["&"]) let splitter = splitter.new(["&"])

View File

@@ -1,6 +1,6 @@
import gleam/bool import gleam/bool
import gleam/int import gleam/int
import gleam/list import gleam/list.{Continue, Stop}
import gleam/option.{type Option, None, Some} import gleam/option.{type Option, None, Some}
import gleam/result import gleam/result
@@ -101,6 +101,49 @@ pub fn try_parsers(
} }
} }
pub fn parse_min_max(str, min, max, parse_fn) {
use <- bool.guard(when: min < 0 || max <= 0 || min > max, return: Error(Nil))
case
list.repeat("", max)
|> list.fold_until(Ok(#("", str, 0)), fn(acc, _) {
let assert Ok(#(hex, str, i)) = acc
case parse_fn(str) {
Error(_) ->
case i < min {
True -> Stop(Error(Nil))
False -> Stop(Ok(#(hex, str, i)))
}
Ok(#(l, rest)) -> Continue(Ok(#(hex <> l, rest, i + 1)))
}
})
{
Error(_) -> Error(Nil)
Ok(#(hex, str, _)) -> Ok(#(hex, str))
}
}
pub fn parse_optional(str, opt_fn) {
case opt_fn(str) {
Error(Nil) -> Ok(#("", str))
Ok(r) -> Ok(r)
}
}
pub fn parse_this_then(
parsers: List(fn(String) -> Result(#(String, String), Nil)),
str: String,
) {
list.fold_until(parsers, Ok(#("", str)), fn(acc, parser) {
let assert Ok(#(res, str)) = acc
case parser(str) {
Ok(#(res2, rest)) -> {
Continue(Ok(#(res <> res2, rest)))
}
Error(Nil) -> Stop(Error(Nil))
}
})
}
pub fn get_multiple( pub fn get_multiple(
to_run: fn(String) -> Result(#(String, String), Nil), to_run: fn(String) -> Result(#(String, String), Nil),
str: String, str: String,
@@ -137,6 +180,41 @@ fn do_get_multiple(
} }
} }
pub fn combine_uris(uris: List(Uri)) -> Uri {
list.fold(uris, Uri(None, None, None, None, "", None, None), fn(acc, uri) {
let acc = case uri {
Uri(Some(scheme), _, _, _, _, _, _) -> Uri(..acc, scheme: Some(scheme))
_ -> acc
}
let acc = case uri {
Uri(_, Some(userinfo), _, _, _, _, _) ->
Uri(..acc, userinfo: Some(userinfo))
_ -> acc
}
let acc = case uri {
Uri(_, _, Some(host), _, _, _, _) -> Uri(..acc, host: Some(host))
_ -> acc
}
let acc = case uri {
Uri(_, _, _, Some(port), _, _, _) -> Uri(..acc, port: Some(port))
_ -> acc
}
let acc = case uri {
Uri(_, _, _, _, path, _, _) if path != "" -> Uri(..acc, path: path)
_ -> acc
}
let acc = case uri {
Uri(_, _, _, _, _, Some(query), _) -> Uri(..acc, query: Some(query))
_ -> acc
}
case uri {
Uri(_, _, _, _, _, _, Some(fragment)) ->
Uri(..acc, fragment: Some(fragment))
_ -> acc
}
})
}
pub fn normalise(uri: Uri) -> Uri { pub fn normalise(uri: Uri) -> Uri {
let percent_splitter = splitter.new(["%"]) let percent_splitter = splitter.new(["%"])
let percent_normaliser = normalise_percent(percent_splitter, _) let percent_normaliser = normalise_percent(percent_splitter, _)
@@ -303,6 +381,10 @@ pub fn parse_hex_digit(str) {
} }
} }
pub fn parse_hex_digits(str, min, max) {
parse_min_max(str, min, max, parse_hex_digit)
}
fn encoding_not_needed(i: Int) -> Bool { fn encoding_not_needed(i: Int) -> Bool {
// $-_.+!*'() // $-_.+!*'()
case i { case i {

View File

@@ -90,5 +90,3 @@ RFC 3986 URI Generic Syntax January 2005
gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@" gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@"
sub-delims = "!" / "$" / "&" / "'" / "(" / ")" sub-delims = "!" / "$" / "&" / "'" / "(" / ")"
/ "*" / "+" / "," / ";" / "=" / "*" / "+" / "," / ";" / "="
" # % < > [ \ ] ^ ` { | }