Did you consider letting the API exploit some of the shiny elegant Perl 6 features like assignable (rw) accessors and overloadable subscript operators?

This seems so Perl 5'ish to me:

$request.url('http://example.com/web/service');
$request.add-field(id => 37271);

Whereas Perl 6 allows the following, which would be a much more elegant I think:

$request.url = 'http://example.com/web/service';
$request.fields<id> = 37271;

URLs are now outdated...

[HTTP::Client] Failed to resolve host name 'http.perl6.org' with family 0.

Trying to upload a binary encoded file currently fails due to a bug in lib/HTTP/Client/Request.pm6 on line 218. The error given is

Cannot use a Buf as a string, but you called the Stringy method on it

The code I used to call it was as follows:

$request.add-file(
	:name($distname),
	:filename($distname),
	:type("application/gzip")
	:content($dist.IO.slurp(:bin))
	:binary
);

I've discussed this in #perl6 for a bit, and gotten some information there. It might be helpful to you.

right, but it's concatenating the supplied binary data (the POST's content) and a final CRLF into a Str. which will treat it as Unicode in NFG and do things like trying to normalize it

yeh, it's building an http request in a Str. I'd argue that's wrong from the very start; it should be using a buf8, any Str-s put into it should be encoded as specified by encoding headers or otherwise .. pedantically ought to reject but practically likely needs to use iso8859-1

Any attempt to access the contents of an HTTP Response results in a crash, with an exception being thrown.

Exception:

Nominal type check failed for parameter '$str'; expected Str but got Array instead
in sub unbase at gen/parrot/CORE.setting:6833
in block at lib/HTTP/Client/Response.pm6:113
in method dechunk at lib/HTTP/Client/Response.pm6:106
in method contents at lib/HTTP/Client/Response.pm6:140
in method content at lib/HTTP/Client/Response.pm6:146
in block at test.p6:7

Full source being used to test:
1 use v6;
2 use HTTP::Client;
3
4 my $client = HTTP::Client.new;
5 my $response = $client.get('http://www.whatismyip.com/');
6 if $response.success {
7 say $response.content;
8 }

It seems like it does not support https. The error I'm getting is pretty straightforward:

Unsupported protocol, 'https'.

Such a pity!

For those who are wondering if there is any other way, try HTTP::UserAgent or LWP::Simple.

Just change the whole thing.

Hi

I'm tracking progress of 2.0 version (to use it later in JSON::RPC::Client as you suggested) and I see that current version has the same performance issue as LWP::Simple - it is parsing URL on every request. In most API services URL is constant and only body changes between calls, so reparsing URL is a waste of CPU.

My idea is to allow passing URI object to Request object to skip parsing. Either by adding uri( ) method or overloading url( ) method.

For example:

multi method url ( URI $uri ) 
{
  $!uri = $uri;
  my $authbit = $!uri.grammar.parse_result<URI_reference><URI><authority>;
  if $authbit<userinfo>
   ... # rest of current url method
}

multi method url ( Str $url ) {
    self.url( URI.new( $url ) )
}

Without running the internal HTTP::Easy example, i cannot depend on HTTP::Client. Could you explain what the issue you are having with rakudo perl6? I want to use HTTP::Client for https://github.com/azawawi/farabi6/

Even if i install HTTP::Client without executing any tests, i still get "Failed to get $url":

use HTTP::Client;
my $client = HTTP::Client.new;
my $url = 'http://google.com';
my $response = $client.get($url);
if ($response.success) {
        say "Got the following: $url";
} else {
        say "Failed while getting $url";
}

Change references of Perl 6 to Raku

The following program called mine returns the following error :

Type check failed in binding $str; expected 'Str' but got 'Array'
in sub unbase at src/gen/m-CORE.setting:6857
in block at lib/HTTP/Client/Response.pm6:113
in method dechunk at lib/HTTP/Client/Response.pm6:106
in method contents at lib/HTTP/Client/Response.pm6:140
in method content at lib/HTTP/Client/Response.pm6:146
in sub slurp at mine:9
in block at mine:15

use HTTP::Client;
proto sub slurp(|) { * }

multi sub slurp($filename, :$bin = False, :$enc = 'utf8') {
if $filename.index: '://' -> $i {

require HTTP::Client;

   my $client = HTTP::Client.new;
   my $response = $client.get($filename);
   return  $response.content if $response.success ;
   fail;
} 
$filename.IO.slurp(:$bin, :$enc);

}
say slurp("http://paris.mongueurs.net");

For info, the page acceded with wget with the header printed

wget -S mongueurs.net
--2014-06-16 11:48:03-- http://mongueurs.net/
Resolving mongueurs.net... 91.216.186.189
Connecting to mongueurs.net|91.216.186.189|:80... connected.
HTTP request sent, awaiting response...
HTTP/1.1 200 OK
Server: Apache/1.3.42 (Unix)
Last-Modified: Mon, 16 Jun 2014 09:45:05 GMT
ETag: "1bc3f8b-3e95-539ebca1"
Content-Type: text/html; charset=UTF-8
Transfer-Encoding: chunked
Date: Mon, 16 Jun 2014 09:48:01 GMT
X-Varnish: 768907453
Age: 0
Via: 1.1 varnish
Connection: keep-alive
Length: unspecified [text/html]
Saving to: ‘index.html.2’

[  <=>                                  ] 16,021      70.7KB/s   in 0.2s   

2014-06-16 11:48:04 (70.7 KB/s) - ‘index.html.’ saved [16021]