Polarion Software Weblog

The described scenario assumes you want to open Polarion ALM to external customers or suppliers

If you are just running a Polarion ALM instance that is directly accessible from the Internet you also expose your Subversion repository , confidential content is accessible from outside.

Actually the Subversion repository is already protected by the access file, but it is easy to expose content by mistake.

This blog describes an approach to open Polarion ALM to external customers and avoid the risk described above.

Internal users access Polarion ALM directly from inside the internal network. The change will be transparent for them.

External users access Polarion ALM via a proxy provided by Apache. This proxy will forward safe requests only which don’t reveal internal content of your repository, even if Subversion’s access file is not configured to hide that information.

Basic configuration of the proxy
The basic configuration of the Apache running the proxy is quite simple.

Adding the following lines to the default configuration is sufficient (replace <polarion-server> with the URL of your Polarion server):
LoadModule proxy_module modules/mod_proxy.so
LoadModule proxy_http_module modules/mod_proxy_http.so
ProxyRequests Off
ProxyPass /polarion <polarion-server>/polarion
ProxyPassReverse /polarion <polarion-server>/polarion

With this configuration all requests to <external-server>/polarion will be passed to <internal-server>/polarion. Please make sure that the Firewall grants access from external-server to internal-server on port 80.

Securing the Proxy
With this configuration an external user could still access files in the repository. Actually, it is not that bad, as the user has to know the exact path of the file he tries to access.

Following line closes this possibility:
ProxyPass /polarion/webdav !

Unfortunately this configuration has a disadvantage: It will also disable access to LiveDocuments using Word and Excel for the external users.

Re-enabling access to Attachment
Disabling all access to the repository also prevents users from downloading attachments.

To enable download of attachments, the /svnwebclient/fileDownload.jsp URL has to be forwarded to the internal server.

Add following lines to the configuration of your Apache acting as proxy:
ProxyPass /svnwebclient/fileDownload.jsp <polarion-server>/svnwebclient/fileDownload.jsp
ProxyPassReverse /svnwebclient/fileDownload.jsp <polarion-server>/svnwebclient/fileDownload.jsp

Doing that without any additional precautions is risky: Again, the access of the external user to files in the repository is only restricted by the access configuration.

An additional layer of security can be implemented by restricting the parameters of the URL to only allow download of files inside the project(s) the external user is able to access. This is achieved by using mod_rewrite to redirect every forbidden access to a nonexistent URL.

Add following lines to your configuration:
LoadModule rewrite_module modules/mod_rewrite.so
RewriteEngine On
RewriteCond %{REQUEST_URI} ^/svnwebclient/fileDownload.jsp$
RewriteCond %{QUERY_STRING} !^.*&url=<project-location>
RewriteRule .* /nonexistent

<project-location> has to be replaced with the location of the project in the repository, all slashes (/) in have to be replaced with %2F and a %2F should be appended at the end.

Example:
Assuming you want to expose Library project from the Polarion demo-data and your internal Polarion server has the URL http://polarion.

You can get the location of the project from the Overview Topic of the project.

The location is Demo%20Projects/demolibrary, the string to use is Demo%20Projects%2Fdemolibrary%2F

The resulting configuration to use for the proxy is:

LoadModule proxy_module modules/mod_proxy.so
LoadModule proxy_http_module modules/mod_proxy_http.so
ProxyRequests Off
ProxyPass /polarion http://polarion/polarion
ProxyPassReverse /polarion http://polarion/polarion
ProxyPass /polarion/webdav !
ProxyPass /svnwebclient/fileDownload.jsp http://polarion/svnwebclient/fileDownload.jsp
ProxyPassReverse /svnwebclient/fileDownload.jsp http://polarion/svnwebclient/fileDownload.jsp
LoadModule rewrite_module modules/mod_rewrite.so
RewriteEngine On
RewriteCond %{REQUEST_URI} ^/svnwebclient/fileDownload.jsp$
RewriteCond %{QUERY_STRING} !^.*&url=Demo%20Projects%2Fdemolibrary%2F
RewriteRule .* /nonexistent

Best Wishes
Benjamin

The following statement will enable log rotation for Apache. The command is specified for installation in a Windows system. Please check paths for you own installation.
CustomLog “|C:/Polarion3.1.1/bundled/apache_2.0.59/bin/rotatelogs.exe C:/Polarion3.1.1/data/logs/apache/access.log 60″ common

Log rotation works by a pipe to the rotatelogs.exe, which itself can be configured to rotate logs by time or by size. Try rotatelogs -? to get help.

The given statement is suitable for the access log file, which receive most of the log output. Look for further log configuration in httpd.conf and adopt them appropriately to rotate all Apache log files.

Best Wishes
Matthias

SUPPORT BULLETIN
Since at least v. 2.0.1 Polarion has had a mistake in the default workflow configuration. In the Reopen action, one of the parameters of a workflow function (valid.states) has a status value "opened" as shown in the following code example:

<action id="reopen" name="Reopen">
    <cleared>
        <field name="resolution"/>
        <field name="resolvedOn"/>
    </cleared>

    <functions>
        <function name="LinkedWorkItemsStatusChange">
            <param name="back.link.roles" value="depends_on"/>
            <param name="link.roles" value="parent"/>
            <param name="valid.states" value="opened,inprogress,reopened"/>
            <param name="target.state" value="reopened"/>
            <param name="clear.attribs" value="resolution,resolvedOn"/>
        </function>
    </functions>
</action>

The correct word is “open”, and this has been changed in version 3.1. If you are upgrading to that version and have an existing workflow configuration form any older version, you should check both you global configuration and any project configurations that may have been based on the default global configuration (which contained the error), and fix the configuration(s) as follows:

<action id="reopen" name="Reopen">
    <cleared>
        <field name="resolution"/>
        <field name="resolvedOn"/>
    </cleared>

    <functions>
        <function name="LinkedWorkItemsStatusChange">
            <param name="back.link.roles" value="depends_on"/>
            <param name="link.roles" value="parent"/>
            <param name="valid.states" value="open,inprogress,reopened"/>
            <param name="target.state" value="reopened"/>
            <param name="clear.attribs" value="resolution,resolvedOn"/>
        </function>
    </functions>
</action>

Polarion 3.0.3 Release is out. The new release ships with a hidden but quite interesting improvement of its wiki functionality.
The feature is called “embedding of workitems inside wiki pages”. Instead of simple references you can directly display workitems inside your wiki pages.

In this blog I want to give you some examples how this functionality can come very handy inside your projects.
Just by embedding simple workitem queries in your wiki pages you will be able to create custom specification documents, impact reports or project status reports.

Lets have a look at the different ways how work items can be embedded.

Embedding via Workitem ID

Use this if you want to embed exactly one workitem inside your wiki page.
Wiki syntax:
{wi: ProjectName/WorkItemId | List of Fields | expand=yes/no}

ProjectName(optional): name of project where work item is located. If no one is specified current project name used by default
WorkItemId: identifier of work item.
fields = List of Fields(optional): list of entries in form FieldName as Type that specifies fields to output. Type can be text, image, image-text. By default the application outputs following fields: id as text, type as image, status as image, severity as image.
expand: if this property is set to yes all specified information about work item is presented below the link

Example:
{wi: MyProject/WI-245 | description as text, status as text-image, severity as text | expand = no}

Embedding By Query

Use this if you want to add multiple items inside your wiki page.
The syntax text is similiar to Polarion Query builder syntax
{wi: query=Polarion Query | List of Fields | sortby = Field tablewidth=Width | tableheight = Height | expand=yes/no}

Example:
{wi: query= type:defect AND (priority:critical OR priority:major) | sortby=created | fields = ID, status as image, severity as image-text, description as text | tablewidth=90% | tableheight=240px}

Creation of custom specification documents

In many cases specification documents consist of some static sections and sections in which you list requirements.

In following Example I combined both approaches - embedding via workitem ID and embedding by query - to create a simple specification document

After some static section at the beginning we list the description of workitem DEMO-1.
Using list as output format( output=list) will display all selected fields below the workitem. I prefer that format because I find it easier to read.
Expand is set to yes (expand=yes) so all listed fields (here only description) are expanded already when you open the wiki page.
We could add additional fields inside the fields section seperated by comma (fields = title as text, description as text…)

In the next section I have added multiple items by query. We have two options here. We can specify a query to filter for specific requirements to be displayed..
For example we could query for all functional requirements with following query: categories.id:functional or
for all requirements in status proposed with following query: status:proposed.
The other option is to pick a set of items by id inside our query.
I have chosen the latter approach in our example.

Displaying linked testcases to a requirement in a table view

You can easily query for linked testcases to a requirement if you know the id(s) of the requirement.
Lets say our requirement has the id DEMO-22 and we want to find all testcases linked to it.
To display requirement DEMO-22 we simply embed it by ID. To display linked testcases in the next column we use following trick.
All linked testcases will have a link to DEMO-22. So if we don’t search by id (id:DEMO-22) but just for the word DEMO-22 we will find all items with the word DEMO-22 inside

.
All testcases that link to DEMO-22 will have the word inside the linked work items section.
As we want to list testcases only the complete query will look like this: “type:testcase AND DEMO-22″

Creating Backlog Reports by User

If you want to embed all open items per user for a specific timpoint you can do that simply by following query:
status:open AND timePoint.id:Iter1 AND assignee.id:ron

Manage Project reports

Create a wiki page that displays items of type project report. The project report item contains personal judgement of the project manager about his project
In that way you can aggregate reports of different projects within one wiki page

Manage risks

Some requirements should be controlled via risks that are assigned to users for periodical review.
Create a risk page which embeds the risks sorted by different aspects

I hope the different scenarios gave you some ideas how that feature could be applied within your project.
Think about: News pages, project announcements, testplan reports, release notes, glossaries
Stay tuned for upcoming blogs.

Best
Tim

In the last blogs we spent quite some time to discuss how you plan items and how you can react on changes in your plan. However we did not use links in our plan. As many items have relationships it is important to understand what impact these relationships can have on your plan.

In Polarion we can find three categories of links; dependency links, parent links, information links.

Dependency Links (depends on)

As already mentioned in our previous blog you should use this type of link if you want to express that one item can only start when the other item has been finished. Typically you will have that type of relationship between tasks. The link should always point from a task to the task it depends on.

Lets have a look at our plan before we will create a dependency relationship.

In this example we will assume that work on “feature 4” can’t start before “feature 1” has been finished. Currently our plan does not reflect that constraint. To achieve this behaviour we have to create a dependency link from “feature 4” to “feature 1”. We edit “feature 4” and add in the field linked workitems “feature 1”. As relationship we select “depends on”. As long as the change is not submitted you will see the workitem id in the title section. After submittion the title will be displayed in the linked workitems section.

We will manually update the plan to see what has changed. As expected “feature 4” has moved now a bit to the right as it has to wait for “feature 2” to be finished. Unfortunately “feature 4” is now also in delay due to the shift by the created dependency

As you can see dependencies are a valuable instrument to adjust and refine your plans. Knowing this people tend to make extensive use of dependency links even if no real “depends on” relationship exist between the items. They use the link more to influence the order in which the items are planned. Don’t do this - use priorities instead

Parent Links (parent, implements)

Imagine estimated a feature with an effort of 10 days. To get a better understanding of the feature the assignee breaks id down to two subfeatures: “feature 4a” and “feature 4b”. After refining the feature and describing it in more detail you add new estimates to “feature 4a” and “feaeture 4b”. No imagine that you estimate for “feature 4a” an estimate of 6 days and to “feature 4b” an estimate of 8 days. We have created a conflict in our plan. The high level feature has been estimated with a different value than its refinement.

Whenever you use parent links polarion will plan the child items and not the parent item. That means that in our liveplan we will from that point on only see the child items.

As you can see the liveplan always displays the highest refinement level. Probably you ask yourself now what happens with the estimates of the parent workitem and where can you control the status of the parent element? Whenever we want to have a view on the high level items we can consult the roadmap view which contains all items from a timepoint that you select. Personally I prefer an adjusted treeview in which I can also see why an item is in delay and what child item coused the delay.

In our example “feature 4a” is in delay and therefore “feature 4” is in delay. If the entry in the timePoint column is in red color it is in delay. “Feature 4” can only be finished at 24^th of April which is exactly 4 days later than initially planned.

Information Links (relates to, duplicates, reveals)

Information links are the easiest links related to planning as they have no impact at all. You can create as many information links as you like, they will have not change your plan. Inks are anyhow helpful to describe that some items have relationships to each other that should be taken in account. Examples are: A bug is related to a feature, a testcase reveals a bug, a change request duplicates a change request that exists already in the system.

Best Wishes
Tim