view doc/xmlrpc.txt @ 5087:39af8a0f3446

Applied patch attached to issue2550723. Problem in index page @pagesize=0 not properly propigated to previous batch link (is set to actual number of entries displayed). Using demo.py verified original issue with 0 page size. Also verified that a non-zero page size was properly propigated (values 5 and 20). Applied patch attached to issue. Verified issue with size 0 fixed. Verified that values 5 and 20 worked correctly. Updated CHANGES.txt.
author John Rouillard <rouilj@ieee.org>
date Sat, 18 Jun 2016 23:49:41 -0400
parents 47cc50617e19
children ade4bbc2716d
line wrap: on
line source

=========================
XML-RPC access to Roundup
=========================

.. contents::
   :local:

Introduction
------------
Version 1.4 of Roundup includes an XML-RPC frontend for remote access. The
XML-RPC interface allows a limited subset of commands similar to those found in
local `roundup-admin` tool.

By default XML-RPC is accessible from ``/xmlrpc`` endpoint:

    http://username:password@localhost:8000/xmlrpc

For demo tracker the URL would be:

    http://localhost:8917/demo/xmlrpc

Enabling XML-RPC server 
-----------------------
There are two ways to run the XML-RPC interface:

  stand alone roundup-xmlrpc-server

  through roundup itself

stand alone roundup-xmlrpc-server
---------------------------------
The Roundup XML-RPC standalone server must be started before remote clients can access the
tracker via XML-RPC. ``roundup-xmlrpc-server`` is installed in the scripts
directory alongside ``roundup-server`` and roundup-admin``. When invoked, the
location of the tracker instance must be specified.

	roundup-xmlrpc-server -i ``/path/to/tracker``

The default port is ``8000``. An alternative port can be specified with the
``--port`` switch.

through roundup
---------------
In addition to running a stand alone server described above, the
XML-RPC service is available from the roundup HTTP server.

security consideration
----------------------
Note that the current ``roundup-xmlrpc-server`` implementation does not
support SSL. This means that usernames and passwords will be passed in
cleartext unless the server is being proxied behind another server (such as
Apache or lighttpd) that provide SSL.


Client API
----------
The server currently implements four methods. Each method requires that the
user provide a username and password in the HTTP authorization header in order
to authenticate the request against the tracker.

======= ====================================================================
Command Description
======= ====================================================================
schema

        Fetch tracker schema.

list    arguments: *classname, [property_name]*

        List all elements of a given ``classname``. If ``property_name`` is
        specified, that is the property that will be displayed for each
        element. If ``property_name`` is not specified the default label
        property will be used.

display arguments: *designator, [property_1, ..., property_N]*

        Display a single item in the tracker as specified by ``designator``
        (e.g. issue20 or user5). The default is to display all properties
        for the item. Alternatively, a list of properties to display can be
        specified.

create  arguments: *classname, arg_1 ... arg_N*

        Create a new instance of ``classname`` with ``arg_1`` through
        ``arg_N`` as the values of the new instance. The arguments are
        name=value pairs (e.g. ``status='3'``).

set     arguments: *designator, arg_1 ... arg_N*

        Set the values of an existing item in the tracker as specified by
        ``designator``. The new values are specified in ``arg_1`` through
        ``arg_N``. The arguments are name=value pairs (e.g. ``status='3'``).

lookup  arguments: *classname, key_value*

        looks up the key_value for the given class. The class needs to
        have a key and the user needs search permission on the key
        attribute and id for the given classname.

filter  arguments: *classname, list or None, attributes*
        
        ``list`` is a list of ids to filter. It can be set to None to run
        filter over all values (requires ``allow_none=True`` when
        instantiating the ServerProxy). The ``attributes`` are given as a 
        dictionary of name value pairs to search for. See also :ref:`query-tracker`.
======= ====================================================================

sample python client
====================
::

        >>> import xmlrpclib
        >>> roundup_server = xmlrpclib.ServerProxy('http://admin:admin@localhost:8917/demo/xmlrpc', allow_none=True)
        >>> roundup_server.schema()
        {'user': [['username', '<roundup.hyperdb.String>'], ...], 'issue': [...]}
        >>> roundup_server.list('user')
        ['admin', 'anonymous', 'demo']
        >>> roundup_server.list('issue', 'id')
        ['1']
        >>> roundup_server.display('issue1')
        {'assignedto' : None, 'files' : [], 'title' = 'yes, ..... }
        >>> roundup_server.display('issue1', 'priority', 'status')
        {'priority' : '1', 'status' : '2'}
        >>> roundup_server.set('issue1', 'status=3')
        >>> roundup_server.display('issue1', 'status')
        {'status' : '3' }
        >>> roundup_server.create('issue', "title='another bug'", "status=2")
        '2'
        >>> roundup_server.filter('user',None,{'username':'adm'})
        ['1']
        >>> roundup_server.filter('user',['1','2'],{'username':'adm'})
        ['1']
        >>> roundup_server.filter('user',['2'],{'username':'adm'})
        []
        >>> roundup_server.filter('user',[],{'username':'adm'})
        []
        >>> roundup_server.lookup('user','admin')
        '1'

Roundup Issue Tracker: http://roundup-tracker.org/