bugzilla-4intranet/Bugzilla/WebService.pm

# -*- Mode: perl; indent-tabs-mode: nil -*-
#
# The contents of this file are subject to the Mozilla Public
# License Version 1.1 (the "License"); you may not use this file
# except in compliance with the License. You may obtain a copy of
# the License at http://www.mozilla.org/MPL/
#
# Software distributed under the License is distributed on an "AS
# IS" basis, WITHOUT WARRANTY OF ANY KIND, either express or
# implied. See the License for the specific language governing
# rights and limitations under the License.
#
# The Original Code is the Bugzilla Bug Tracking System.
#
# Contributor(s): Marc Schumann <wurblzap@gmail.com>
#                 Max Kanat-Alexander <mkanat@bugzilla.org>
#                 Rosie Clarkson <rosie.clarkson@planningportal.gov.uk>
#                 
# Portions © Crown copyright 2009 - Rosie Clarkson (development@planningportal.gov.uk) for the Planning Portal

# This is the base class for $self in WebService method calls. For the 
# actual RPC server, see Bugzilla::WebService::Server and its subclasses.
package Bugzilla::WebService;
use strict;
use Date::Parse;
use XMLRPC::Lite;

sub datetime_format {
    my ($self, $date_string) = @_;

    my $time = str2time($date_string);
    my ($sec, $min, $hour, $mday, $mon, $year) = localtime $time;
    # This format string was stolen from SOAP::Utils->format_datetime,
    # which doesn't work but which has almost the right format string.
    my $iso_datetime = sprintf('%d%02d%02dT%02d:%02d:%02d',
        $year + 1900, $mon + 1, $mday, $hour, $min, $sec);
    return $iso_datetime;
}

# For some methods, we shouldn't call Bugzilla->login before we call them
use constant LOGIN_EXEMPT => { };

sub login_exempt {
    my ($class, $method) = @_;

    return $class->LOGIN_EXEMPT->{$method};
}

sub type {
    my ($self, $type, $value) = @_;
    if ($type eq 'dateTime') {
        $value = $self->datetime_format($value);
    }
    return XMLRPC::Data->type($type)->value($value);
}

1;

__END__

=head1 NAME

Bugzilla::WebService - The Web Service interface to Bugzilla

=head1 DESCRIPTION

This is the standard API for external programs that want to interact
with Bugzilla. It provides various methods in various modules.

Currently the only method of accessing the API is via XML-RPC. The XML-RPC
standard is described here: L<http://www.xmlrpc.com/spec>

The endpoint for Bugzilla WebServices is the C<xmlrpc.cgi> script in
your Bugzilla installation. For example, if your Bugzilla is at
C<bugzilla.yourdomain.com>, then your XML-RPC client would access the
API via: C<http://bugzilla.yourdomain.com/xmlrpc.cgi>

=head1 CALLING METHODS

Methods are called in the normal XML-RPC fashion. Bugzilla does not currently
implement any extensions to the standard method of XML-RPC method calling.

Methods are grouped into "packages", like C<Bug> for 
L<Bugzilla::WebService::Bug>. So, for example,
L<Bugzilla::WebService::Bug/get>, is called as C<Bug.get> in XML-RPC.

=head1 PARAMETERS

In addition to the standard parameter types like C<int>, C<string>, etc.,
XML-RPC has two data structures, a C<< <struct> >> and an C<< <array> >>.

=head2 Structs

In Perl, we call a C<< <struct> >> a "hash" or a "hashref". You may see
us refer to it that way in the API documentation.

In example code, you will see the characters C<{> and C<}> used to represent
the beginning and end of structs.

For example, here's a struct in XML-RPC:

 <struct>
   <member>
     <name>fruit</name>
     <value><string>oranges</string></value>
   </member>
   <member>
     <name>vegetable</name>
     <value><string>lettuce</string></value>
   </member>
 </struct>

In our example code in these API docs, that would look like:

 { fruit => 'oranges', vegetable => 'lettuce' }

=head2 Arrays

In example code, you will see the characters C<[> and C<]> used to
represent the beginning and end of arrays.

For example, here's an array in XML-RPC:

 <array>
   <data>
     <value><i4>1</i4></value>
     <value><i4>2</i4></value>
     <value><i4>3</i4></value>
   </data>
 </array>

In our example code in these API docs, that would look like:

 [1, 2, 3]

=head2 How Bugzilla WebService Methods Take Parameters

B<All> Bugzilla WebServices functions take their parameters in
a C<< <struct> >>. Another way of saying this would be: All functions
take a single argument, a C<< <struct> >> that contains all parameters.
The names of the parameters listed in the API docs for each function are
the C<name> element for the struct C<member>s.

=head1 LOGGING IN

You can use L<Bugzilla::WebService::User/login> to log in as a Bugzilla 
user. This issues standard HTTP cookies that you must then use in future
calls, so your XML-RPC client must be capable of receiving and transmitting
cookies.

=head1 STABLE, EXPERIMENTAL, and UNSTABLE

Methods are marked B<STABLE> if you can expect their parameters and
return values not to change between versions of Bugzilla. You are 
best off always using methods marked B<STABLE>. We may add parameters
and additional items to the return values, but your old code will
always continue to work with any new changes we make. If we ever break
a B<STABLE> interface, we'll post a big notice in the Release Notes,
and it will only happen during a major new release.

Methods (or parts of methods) are marked B<EXPERIMENTAL> if 
we I<believe> they will be stable, but there's a slight chance that 
small parts will change in the future.

Certain parts of a method's description may be marked as B<UNSTABLE>,
in which case those parts are not guaranteed to stay the same between
Bugzilla versions.

=head1 ERRORS

If a particular webservice call fails, it will throw a standard XML-RPC
error. There will be a numeric error code, and then the description
field will contain descriptive text of the error. Each error that Bugzilla
can throw has a specific code that will not change between versions of
Bugzilla.

The various errors that functions can throw are specified by the
documentation of those functions.

If your code needs to know what error Bugzilla threw, use the numeric
code. Don't try to parse the description, because that may change
from version to version of Bugzilla.

Note that if you display the error to the user in an HTML program, make
sure that you properly escape the error, as it will not be HTML-escaped.

=head2 Transient vs. Fatal Errors

If the error code is a number greater than 0, the error is considered
"transient," which means that it was an error made by the user, not
some problem with Bugzilla itself.

If the error code is a number less than 0, the error is "fatal," which
means that it's some error in Bugzilla itself that probably requires
administrative attention.

Negative numbers and positive numbers don't overlap. That is, if there's
an error 302, there won't be an error -302.

=head2 Unknown Errors

Sometimes a function will throw an error that doesn't have a specific
error code. In this case, the code will be C<-32000> if it's a "fatal"
error, and C<32000> if it's a "transient" error.

=head1 COMMON PARAMETERS

Many Webservice methods take similar arguments. Instead of re-writing
the documentation for each method, we document the parameters here, once,
and then refer back to this documentation from the individual methods
where these parameters are used.

=head2 Limiting What Fields Are Returned

Many WebService methods return an array of structs with various
fields in the structs. (For example, L<Bugzilla::WebService::Bug/get>
returns a list of C<bugs> that have fields like C<id>, C<summary>, 
C<creation_time>, etc.)

These parameters allow you to limit what fields are present in
the structs, to possibly improve performance or save some bandwidth.

=over

=item C<include_fields> (array)

An array of strings, representing the (case-sensitive) names of fields.
Only the fields specified in this hash will be returned, the rest will
not be included.

If you specify an empty array, then this function will return empty
hashes.

Invalid field names are ignored.

Example:

  User.get( ids => [1], include_fields => ['id', 'name'] )

would return something like:

  { users => [{ id => 1, name => 'user@domain.com' }] }

=item C<exclude_fields> (array)

An array of strings, representing the (case-sensitive) names of fields.
The fields specified will not be included in the returned hashes.

If you specify all the fields, then this function will return empty
hashes.

Invalid field names are ignored.

Specifying fields here overrides C<include_fields>, so if you specify a
field in both, it will be excluded, not included.

Example:

  User.get( ids => [1], exclude_fields => ['name'] )

would return something like:

  { users => [{ id => 1, real_name => 'John Smith' }] }

=back


=head1 EXTENSIONS TO THE XML-RPC STANDARD

=head2 Undefined Values

Normally, XML-RPC does not allow empty values for C<int>, C<double>, or
C<dateTime.iso8601> fields. Bugzilla does--it treats empty values as
C<undef> (called C<NULL> or C<None> in some programming languages).

Bugzilla also accepts an element called C<< <nil> >>, as specified by 
the XML-RPC extension here: L<http://ontosys.com/xml-rpc/extensions.php>, 
which is always considered to be C<undef>, no matter what it contains.

Bugzilla does not use C<< <nil> >> values in returned data, because currently
most clients do not support C<< <nil> >>. Instead, any fields with C<undef>
values will be stripped from the response completely. Therefore
B<the client must handle the fact that some expected fields may not be 
returned>.

=begin private

nil is implemented by XMLRPC::Lite, in XMLRPC::Deserializer::decode_value
in the CPAN SVN since 14th Dec 2008 
L<http://rt.cpan.org/Public/Bug/Display.html?id=20569> and in Fedora's 
perl-SOAP-Lite package in versions 0.68-1 and above.

=end private