===========
Roundup FAQ
===========

.. contents::
   :local:


Installation
------------

Living without a mailserver
~~~~~~~~~~~~~~~~~~~~~~~~~~~

Remove the nosy reactor - delete the tracker file
``detectors/nosyreactor.py`` from your tracker home.


The cgi-bin is very slow!
~~~~~~~~~~~~~~~~~~~~~~~~~

Yep, it sure is. It has to start up Python and load all of the support
libraries for *every* request.

The solution is to use the built in server (or possibly the mod_python
or WSGI support).

To make Roundup more seamless with your website, you may place the built
in server behind apache and link it into your web tree (see below).


How do I put Roundup behind Apache
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

We have a project (foo) running on ``tracker.example:8080``.
We want ``http://tracker.example/issues`` to use the roundup server, so we 
set that up on port 8080 on ``tracker.example`` with the ``config.ini`` line::

  [tracker]
  ...
  web = 'http://tracker.example/issues/'

We have a "foo_issues" tracker and we run the server with::

  roundup-server -p 8080 issues=/home/roundup/trackers/issues 

Then, on the Apache machine (eg. redhat 7.3 with apache 1.3), in
``/etc/httpd/conf/httpd.conf`` uncomment::

  LoadModule proxy_module       modules/libproxy.so

and::

  AddModule mod_proxy.c

Then add::

  # roundup stuff (added manually)
  <IfModule mod_proxy.c>
  # proxy through one tracker
  ProxyPass /issues/ http://tracker.example:8080/issues/
  # proxy through all tracker(*)
  #ProxyPass /roundup/ http://tracker.example:8080/
  </IfModule>

Then restart Apache. Now Apache will proxy the request on to the
roundup-server.

Note that if you're proxying multiple trackers, you'll need to use the
second ProxyPass rule described above. It will mean that your TRACKER_WEB
will change to::

  TRACKER_WEB = 'http://tracker.example/roundup/issues/'

Once you're done, you can firewall off port 8080 from the rest of the world.

Note that in some situations (eg. virtual hosting) you might need to use a
more complex rewrite rule instead of the simpler ProxyPass above. The
following should be useful as a starting template::

  # roundup stuff (added manually)
  <IfModule mod_proxy.c>

  RewriteEngine on
  
  # General Roundup
  RewriteRule ^/roundup$  roundup/    [R]
  RewriteRule ^/roundup/(.*)$ http://tracker.example:8080/$1   [P,L]
  
  # Handle Foo Issues
  RewriteRule ^/issues$  issues/    [R]
  RewriteRule ^/issues/(.*)$ http://tracker.example:8080/issues/$1 [P,L]
  
  </IfModule>


How do I run Roundup through SSL (HTTPS)?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

The preferred way of using SSL is to proxy through apache and use its
SSL service. See the previous question on how to proxy through apache.

The standalone roundup-server now also has SSL support which is still
considered experimental. For details refer to the documentation of
roundup server, in particular to the generated configuration file
generated with ::

    roundup-server --save-config

that describes the needed option in detail. With the standalone server
now XMLRPC over SSL works, too.


Roundup runs very slowly on my XP machine when accessed from the Internet
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

The issue is probably related to host name resolution for the client
performing the request. You can turn off the resolution of the names
when it's so slow like this. To do so, edit the module
roundup/scripts/roundup_server.py around line 77 to add the following
to the RoundupRequestHandler class::

     def address_string(self):
         return self.client_address[0]


Templates
---------

What is that stuff in the tracker html directory?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

This is the template code that Roundup uses to display the various pages.
This is based upon the template markup language in Zope called, oddly
enough, "Zope Page Templates". There's documentation in the Roundup
customisation_ documentation. For more information have a look at:

   http://docs.zope.org/zope2/zope2book/

specifically chapter 10 "Using Zope Page Templates" and chapter 14 "Advanced
Page Templates".


But I just want a select/option list for ....
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Really easy... edit ``html/issue.item.html``. For ``nosy``, change the line
(around line 69) from::

  <span tal:replace="structure context/nosy/field" />

to::

  <span tal:replace="structure context/nosy/menu" />

For ``assigned to``, this is already done around line 77::

  <td tal:content="structure context/assignedto/menu">assignedto menu</td>



Great! But now the select/option list is too big
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

That is a little harder (but only a little ;^)

Again, edit ``html/issue.item``. For nosy, change line (around line 69) from::

  <span tal:replace="structure context/nosy/field" />

to::

  <span tal:replace="structure python:context.nosy.menu(height=3)" />

for more information, go and read about Zope Page Templates.


Using Roundup
-------------

I got an error and I cannot reload it!
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

If you're using Netscape/Mozilla, try holding shift and pressing reload.
If you're using IE then install Mozilla and try again ;^)


I keep getting logged out
~~~~~~~~~~~~~~~~~~~~~~~~~

Make sure that the ``tracker`` -> ``web`` setting in your tracker's
config.ini is set to the URL of the tracker.

I'm getting infinite redirects in the browser
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

A wrong value for the ``tracker`` -> ``web`` setting may also result in
infinite redirects, see http://issues.roundup-tracker.org/issue2537286


How is sorting performed, and why does it seem to fail sometimes?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

When we sort items in the hyperdb, we use one of a number of methods,
depending on the properties being sorted on:

1. If it's a String, Integer, Number, Date or Interval property, we
   just sort the scalar value of the property. Strings are sorted
   case-sensitively.
2. If it's a Link property, we sort by either the linked item's "order"
   property (if it has one) or the linked item's "id".
3. Mulitlinks sort similar to #2, but we start with the first
   Multilink list item, and if they're the same, we sort by the second item,
   and so on.

Note that if an "order" property is defined on a Class that is used for
sorting, all items of that Class *must* have a value against the "order"
property, or sorting will result in random ordering.

.. _`customisation`: customizing.html