Return to Forge's homepage

Frequently Asked Questions about the INRIA Forge

How to contact INRIA Forge administrators?

Who can start a project on InriaForge?

What should I use the InriaForge for? What is the good usage of the InriaForge?
How to simplify my interface (i.e. remove unuseful tabs)?
What is the default language on InriaForge?
SCM (source control manager) questions
About public, private projects and access control
About mailing lists and InriaForge
About web hosting for projects
About the Wiki plugin How do I setup ssh (even on windows?)
How to configure Eclipse to work with InriaForge?
How to name forums, mailing lists on InriaForge?
About docs area
How can a tracker/forum send mail to a mailing list?
Can a git commit trigger a tracker update?
How are the statistics of the front page computed?
How to change the license for a project?
How to connect Hudson (continuous integration server) to InriaForge?
Miscellaneous other questions or issues
About Spam and InriaForge

How to contact INRIA Forge administrators?

First of all, if you have a question, please check that the answer is not already in this FAQ.

You can submit a support ticket at the inria helpdesk portal (inria account is mandatory). On this portal, select "IT support request", then "ask for a fix" or "service request", and then select "services for research", then "inriaforge".

You can also submit a support ticket by mail at address sed-forge.helpdesk-prc at (this is the way to go if you don't have an inria account).

The old support email address at is an alias to sed-forge.helpdesk-prc at

When submitting your requests, don't forget to provide accurate details of what's going wrong. Read, for example, How to Report Bugs Effectively.

Finally, you can also have a look at our bug tracker.

Who can start a project on InriaForge?

InriaForge has been setup to enable people to work on projects related to Inria. Anyone can take part, but only people working for Inria (i.e using a email,,, or in their profile) can start a project. You can read more about the general terms of service

Who can delete a project?

Deleting a project is a privileged operation that can only be performed by InriaForge's admins. Please sent any request to<AT>

Why can't I find the 'create project' option in my administration page?

Only people working for Inria (i.e using a email,,, or in their profile) can register a project.

How can I remove myself from a project?

To do so :

What should I use the InriaForge for? What is the good usage of the InriaForge?

InriaForge is initially and mainly aimed at providing a set of tools for collaborative work on software development projects. Additionally, it can also be used for collaborative work on projects not directly related to software development, such as collaborative edition of scientific publications, since source code or text edition are similar.

Usually the main InriaForge tools used are source control (git, subversion), but there are also some facilities such as web hosting that can prove useful on their own, but keep in mind that the design and sizing of InriaForge, fusionforge (the software used to run the InriaForge) and their associated services (web hosting, ssh access, mailing lists hosting, etc.) have some limitations:

Hence, some InriaForge usages are rather inappropriate:

Alternative solutions:

The frontier between good and bad usage of the InriaForge is fuzzy, but as there are a lot of users and projects on the forge, if too much people abuse the system we will have to put hard limits and restrictions on the forge usage, which will be annoying for everyone.

How to simplify my interface (i.e. remove unuseful tabs)?

On the admin project page (tab Admin), click on the first link "Edit Public Info".
Now you can select active features like Tracker, File Release System, Wiki, etc ...

What is the default language on InriaForge?

InriaForge's interface is maintained in French and in English. Depending on your browser settings and on your personnal preferences when you are logged in you will see the web interface in French or in English. To be more precise :

Nervertheless, some elements are not bilingual: this FAQ for instance, is only available in English; the news items on the front page are only available in the language the project submitting the news has written them.

Each project also defines a natural language. This is an indication of the language that should be used when submitting bugs or writting to the mailing lists of the project. It has no known influence on the language of the interface, even when browsing the project's pages.

Finally, the websites each project can setup on <project_name> define their own behaviour. The InriaForge team or software has no influence on them.

SCM (source control manager) questions

SCM urls have changed after the June 2, 2015 migration, what can I do?

To accomodate major changes in the user authentication system, the source control repositories urls had to be changed when migrating to fusionforge 6.0 on June 2, 2015. To minimize the burden for users, transparent and automatic relocation to the new urls has been setup for git and subversion repositories. Thus, in the general case, a simple git fetch or svn up will update the repositories urls and users won't even notice the change. There are however situations where automatic relocation does not work, including when using the eclipse subversion client. In these situations you need to:

Can I access the code repository on the InriaForge server?

Your code repository is located on the InriaForge server on host, directory :

It may be accessed directly on the InriaForge server via ssh. The data may also be retrieved using the commands sftp, rsync or scp.
For example, to list the svn code repository on InriaForge server :
    $ sftp <yourlogin><project_name>
    sftp> ls
    sftp> exit

To know who can access the code repository, go to next FAQ entry.

Who can access my code repository?

With the default settings, project members with Doc Writer or Support Tech roles cannot commit changes into the SCM.
Source code repository access depends on the settings of the project and on the privileges of the user trying to access the code repository. The following table sums the situation.


Read access to source code

Write access to source code

Public project/anonymous acces enabled

Anonymous user

Logged user
Project member, no SCM access
Project member, read access to SCM
Project member, read/write access to SCM
Private project or anonymous access disabled Anonymous user
Logged user
Project member, no SCM access
Project member, read access to SCM
Project member, read/write access to SCM

Changing user privileges is done through the edition of user roles in a project's admin tab.

How do I import my Subversion history?

To import your repository, first export a SVN dump:

  svnadmin dump SVN_REPO | gzip > SVN_REPO.dump.gz

Make sure you did upload your SSH key in the web interface and connect to

Then reset your SVN repository (so there's no basic /trunk and /branches directories in SVN, which conflict with the dump).
Caution: you will lose all the data in this SVN repository:

rm -rf /svnroot/<project_name>/db/*
svnadmin create temp/
cp -a temp/db/* /svnroot/<project_name>/db/
rm -rf temp/

Copy your SVN_REPO.dump.gz to the forge (e.g. with scp).

Then import it on the forge:

zcat SVN_REPO.dump.gz | svnadmin load /svnroot/<project_name>/

(Note: this can be long. You may want to use screen.)

Last, fix the permissions so your project members can commit:

  find /svnroot/<project_name>/db/* -type d -print0 | xargs -r -0 chmod 2775
  find /svnroot/<project_name>/db/* -type f -print0 | xargs -r -0 chmod 664
  chgrp -R <project_name>_scmrw /svnroot/<project_name>/db/*

How do I import several repositories in one subversion repository?

First, let's imagine we have three repositories that we want maintained under <project_name>. Let's call them calc, calendar and spreadsheet (names taken from the subverison book). The first step is to create a dump of these repositories and the directories under which these projects will be imported in your repository.
> svnadmin dump -q /path/to/calc/repos > calc-dumpfile
> svnadmin dump -q /path/to/calendar/repos > calendar-dumpfile
> svnadmin dump -q /path/to/spreadsheet/repos > spreadsheet-dumpfile
> svn mkdir -m "Imported project roots" \
     svn+ssh://<yourlogin><project_name>/calc \
     svn+ssh://<yourlogin><project_name>/calendar \
Then, copy the dump file to your home directory on InriaForge
> scp calc-dumpfile  <yourlogin><yourlogin>/
> scp calendar-dumpfile  <yourlogin><yourlogin>/
> scp spreadsheet-dumpfile  <yourlogin><yourlogin>/
Finally, contact the InriaForge admins in order to run such commands
$ svnadmin load /svnroot/<project_name> --parent-dir calc < /home/users/<yourlogin>/calc-dumpfile
$ svnadmin load /svnroot/<project_name> --parent-dir calendar < /home/users/<yourlogin>/calendar-dumpfile
$ svnadmin load /svnroot/<project_name> --parent-dir spreadsheet < /home/users/<yourlogin>/spreadsheet-dumpfile
You now have your three repositories under your project's repository on InriaForge.

I get an error when accessing SVN over SSH, how can I debug?

If you get errors such as:

svn: E170013: Unable to connect to a repository at URL 'svn+ssh://'
svn: E210002: Pour mieux déboguer les problèmes de connexion SSH, retirer l'option -q de 'ssh' dans la section [tunnels] de votre fichier de configuration de Subversion.
svn: E210002: La connexion réseau a été fermée de façon inattendue

svn: E170013: Unable to connect to a repository at URL 'svn+ssh://'
svn: E210002: To better debug SSH connection problems, remove the -q option from 'ssh' in the [tunnels] section of your Subversion configuration file.
svn: E210002: Network connection closed unexpectedly

you can get more information by running e.g.:

SVN_SSH='ssh -v' svn up
OpenSSH_6.9p1 Debian-2, OpenSSL 1.0.2d 9 Jul 2015
debug1: Reading configuration data /home/pro/.ssh/config
debug1: Reading configuration data /etc/ssh/ssh_config

If you can't figure out what is wrong, you can include the output in your ticket to the forge admins.

How can I use git/subversion on windows with ssh authentication?

If you have followed ths configuration steps described in the ssh for windows page mentionned in How do I setup ssh (even on windows?), you can use either a graphical or a shell-based utility to access your git/svn repository.

To use the graphical utility, follow the steps described in our tortoise page.

To use svn from a shell, you can follow these steps:

How can I change the type of SCM used in my project?

In the project page, go in the Admin tab. Then go in the Tools sub-tab, and in the Tool Admin pane, click Source Code Admin. You can then select the kind of source code repository, and click update. This does not destroy the previous project repository, and it is up to the project administrator to manage any synchronization or data migration between the Subversion and Git repositories.

Why can't I no longer download a snapshot of my SCM repository?

Shapshot generation was disabled because it caused to big a load on the server because of the way the snapshot was made on projects with too many branches. There is no generic way of making a snapshot in a way that avoids the problems encoutered. Therefore, projects relying on snapshots should either use the tarball that is generated (and available on the admin page of each project) or make a snapshot themselves.

A snapshot is nothing more than a tarball of the checkout of the complete repository made available each night. You can program the same functionnality whith the script available here for svn.

What version of svn is running on the server?

forge-scm2:~$ date
Wed Jun 17 18:04:23 CEST 2015
forge-scm2:~$ svn --version
svn, version 1.8.10 (r1615264)
   compiled Apr  1 2015, 02:54:56 on x86_64-pc-linux-gnu

Copyright (C) 2014 The Apache Software Foundation.
This software consists of contributions made by many people;
see the NOTICE file for more information.
Subversion is open source software, see

The following repository access (RA) modules are available:

* ra_svn : Module for accessing a repository using the svn network protocol.
  - with Cyrus SASL authentication
  - handles 'svn' scheme
* ra_local : Module for accessing a repository on local disk.
  - handles 'file' scheme
* ra_serf : Module for accessing a repository via WebDAV protocol using serf.
  - using serf 1.3.8
  - handles 'http' scheme
  - handles 'https' scheme

How to configure http proxy with svn webdav?

Edit your subversion configuration file : ~/.subversions/servers and add in global section :

http-proxy-host = <your-proxy-server>
http-proxy-port = <your-proxy-server-port>

How can I add/modify hook script as post-commit for the subversion repository of my project hosted on InriaForge?

A hook script is a program triggered by some repository event, such as a new commit or the modification of an unversioned property.

Hooks are usually managed by the 'scmhook' tool, which can manage a few generic hooks (see FAQ entry for e-mail notifications).

If you are already using customized hooks, do not enable the 'scmhook' tool, or it will overwrite or delete your hooks.

If you have specific needs, contact us to check if they are compliant with our hosting policy, and manually install them in /svnroot/your_project/hooks/ on

For more information concerning the subversion hooks, consult Subversion hook manual page.

How can I receive email notifications for SVN commits or Git pushes (commit email hook)

You (as a project administrator) need to follow the following steps:

  1. Setup a commit mailing list:
  2. Setup a hook on the SCM server:

Make sure to subscribe to the mailing list by filling in Membership Management > Mass subscription from the management webpage of the mailing list.

People who wish to receive the commit mails need to subscribe to the commit mailing list.

The mailing list automatically approves the commit emails, see this entry for details.

How to repair my Subversion repository?

If your subversion repository is corrupted (i.e. messages like "Svn diff too large, svn repository corrupted, ..."), follow this procedure :

This kind of problem can occur when commiting large binary files. A best practice for these files (pdf, doc,code compiled, or release) is to use the feature "file release system" provided by the forge.

How to start with an empty GIT repository?

By default a Git repository is empty and you cannot clone an empty Git repository.
If your Git repository already exists, simply push it. Otherwise, follow this procedure :

mkdir my_project
cd my_project
git init
git remote add origin git+ssh://
echo "Description of your project" > README
git add README # Note: it is mandatory to have a commit to be able to push.
git commit -m "Addition of a readme file"
git push origin master

How do I setup ssh (even on windows?)

To be able to:

you need to setup an ssh connection between your machine and the forge server.

On Linux:

  1. An ssh client should already be installed on your machine and should be accessible from the command line, as the "ssh" command. If not, contact your system administrator.
  2. Generate a pair of rsa public/private keys with the ssh-keygen command: ssh-keygen -t rsa (make sure you enter a lengthy passphrase, ie non null, when asked to).
  3. Copy your public key in an easy-to access file: cp ~/.ssh/ ~.
  4. Paste your public key in the InriaForge website. To do this, you need to go to your account and then go to the Account maintainance tab. At the bottom of the Account maintainance tab, you should see a Shell Account Information section which contains an [Edit keys] link. Paste your public key(s) in the empty field below and click the Update button.

On Windows, consult the ssh for windows page.

A common problem is that the Shell Account Information field does not appear in your account page. This usually happens because you do not belong to any InriaForge project. To correct this problem, either create your own project first or request the project admin of the project you want to join to add you to the list of the project's contributors.

How to migrate from Subversion to GIT?

How to remove some files or rewrite history in a git or subversion repository?

Source control tools such as Subversion and Git are designed to keep track of every operation done on a repository. Thus, usually the right way to revert a commit is to do a new commit which reverts the effect of the first commit. However, in come rare situations this is not appropriate: for example, one may have inadvertently commited a file with sensitive information (such as passwords) or some huge file which cripples the repository. In these situations, the only way to really remove the file is to rewrite the history.

This is an advanced repository operation, that needs to be done by one of the project administrators, and it has some important consequences: all local repositories will need to be re-synchronized. This means all repositories of all project members as well as checkouts of any non members in case of project with anonymous read-only access. The project admin running this operation needs to work in collaboration with other project admins and project members, and should ideally take the repository offline during the time the history is being rewritten (eg. by installing a hook preventing any commit to be done during the maintenance). Also, as there is a risk to completely mess up the server repository (with a manipulation error or in case something goes wrong), please be aware that:

That said, here's how to do that

With Git

There are several ways to do that with git. Here are a few (non exhaustive list):

All working copies then need to be resynchronized.
The simplest way is to re-clone from scratch a new working copy. Or it should also be possible to do:

<user>@localhost:<working copy path>$ git fetch
<user>@localhost:<working copy path>$ git reset --hard origin/master
<user>@localhost:<working copy path>$ git gc --prune=all

With Subversion

Here are the steps to rewrite the history of a subversion project:

All working copies then need to be resynchronized.
As far as we know the only way is to re-checkout from scratch a new working copy.

How to store docker projects?

The storage of docker binary images of docker is discouraged in the forge since:

As a consequence, two solutions in order to share docker images are recommended:

About public, private projects and access control

What is the difference between a public and a private project?

A private project is a project that only exists for those who know that it has been created and that can only be seen by the people that are registered as members of this project. All members of a project have access to the source control Manager (SCM) and to the webspace of the project.

A public project appears in the list of project hosted on the forge. Anybody can find it using the search function, the project tree or a search engine and can browse all public items of the project. But a public project does not necessarily mean that all items of the project are public. The visibility of each item can be controlled item by item. For example, you could have a public tracker for bugs, a private tracker for security problems, and anonymous access to the source. The visibility of each item is controlled by the admin link that is accessible to the project's admins on the page showing the item.

In all cases, only registered members of the project can change the web pages of the project or its source code.

Can a project change visibility (become public?)

Project visibility is a project attribute chosen at the start of the life of a project. Project administrators can change the visibility of their project by changing the properties of the Anonymous role.

This can be done in the project admin pages, in the section "Users and permissions". Set "Project visibility" to "visible" for the Anonymous role to make a project public. If the role does not exist, you should reference (link) the global Anonymous role from your project.

You should be aware that changing the visibility of a project does not change the visibility of the different elements of the forge you are using. After having made a project public, you will need to evaluate if you want to change the visibility of any other element (eg. your SCM repository or your bug tracking tools).

Can I give different access to parts of my project to different users?

The key to controlling access to the different parts of your project on InriaForge is the admin page of your project. On the right of this page, each member of your project is given a role, and to each role, access rights are associated. You can change them by using the edit role part of the page.

Two roles deal with users that are not part of your project: "Logged In" (for registered forge users that do not belong to your project) and "Anonymous" (for all users). These roles can be linked from the administration page to define permissions for external users.

NB : Due to availability of both protocols (https and svn+ssh), restricting access to a Subversion subdirectory for somes users (among those participating in the whole project) is not possible. Creating a private project is the only way to gain this feature.

How can I prevent anonymous users from inserting spam in a tracker or a forum?

Some projects enable anonymous users to post issues on the trackers or forums they run, for reporting for example. Some spammer are using this setting for advertising using the trackers or forums, either by posting new issues, or by commenting open issues.

InriaForge does not have any technical mesures in place to prevent that kind of spam. Therefore, if your project suffers from this, you should change the access rights of all your trackers or forums to prevent anonymous posting, by editing the Anonymous role.

About mailing lists and InriaForge

I cannot access the admin interface for the mailing lists

I created the project and/or I am an admin of a project but my normal InriaForge password does not allow me to access the admin web interface of the mailing lists.

This is normal. The password used to control access to the mailing lists admin web interface is different from your normal InriaForge account. This different password has been emailed to all the admins of the project at the time the mailing-list was created. The Subject of the email was InriaForge New Mailing List. It contained the password used to access to the admin url:

How can I reset a Mailman mailing list administration password?

If you lose/forget the password of the mailing list administrator, you can reset it with the InriaForge interface. On the tab "lists" associated with your project, click on the link "Admin". Then you can reset your mailing list password. The new password will be sent to the adminstrator of the list.

How can I make my mailing lists refuse postings from non-subscribers?

To do this, you need to be the mailing list administrator. Go to: To be able to log successfully, you need your special mailing list administration password. For information on this password, see I cannot access the admin interface for the mailing lists.

  1. Log into the administrative interface of your list
  2. Go to privacy options
  3. Go to the sender filter
  4. Find the Should messages from non-members, which are automatically discarded, be forwarded to the list moderator? line and answer No.
  5. Find the Action to take for postings from non-members for which no explicit action is defined line and answer Discard.
  6. Save your changes

How can I make my mailing lists automatically accept mails from some domains?

To do this, you need to be the mailing list administrator. Go to: To be able to log successfully, you need your special mailing list administration password. For information on this password, see I cannot access the admin interface for the mailing lists.

  1. Log into the administrative interface of your list
  2. Go to privacy options
  3. Go to the sender filter
  4. Find the List of non-member addresses whose postings should be automatically accepted line and add something like (example from the default settings)
  5. Save your changes

How to name mailing lists on InriaForge?

On InriaForge, each mailing list name is of the form :

To know what kind of suffixname is allowed, see this FAQ entry

If you need to add a member to the list that is already registered on InriaForge, you can use the address <userlogin>
This will refer automatically to the email address the user has given when registered on InriaForge.

About web hosting for projects

How can I modify the web hosted pages for my project?

The web hosted pages for a projet on InriaForge are located under host, and directory /home/groups/<project_name>/htdocs/.

By default, a index.php file is created for each project when the project has been registrated on the InriaForge. As you can not access to the server directly through ssh, you must use the commands : sftp, scp, or rsync.

To list the files, you can use sftp:

$ sftp <yourlogin><project_name>/htdocs/
sftp> ls
sftp> exit

In order to modify the web hosted pages, you can use rsync to retrieve the files already hosted at<project_name>/htdocs/ :

$ rsync -avz <yourlogin><project_name>/htdocs/ <localpath>
Note: The parameters proposed below for rsync ensure that symbolic links, devices, attributes, permissions, ownerships, etc. are preserved in the transfer.

Modify the files locally on your machine, then send back the modified pages to<project_name>/htdocs/ :

$ rsync -av <localpath>/ <yourlogin><project_name>/htdocs
If you have trouble with the file permissions, once the files are back to the InriaForge server, you can use sftp to modify them.
$ sftp <yourlogin><project_name>/htdocs/
sftp> ls
sftp> chmod o+r filestobemodified
sftp> exit

Can I have access to the access statistics of my project's webspace?

Yes. You can find an access log file under /home/groups/<project_name>/log and a HTML file containing information generated by awstats under /home/groups/<project_name>/htdocs/log/. You can visualize the generated statistics using the URL :

You could also specify a particular year of statistical data by adding at the end of the url :

Can I implement access control for my project's webspace?

Yes. Apache honors the .htaccess files that you place in the file hierarchy of your projects (files under /home/groups/<project_name>/htdocs/). You can therefore implement access protection using .htpasswd. Please refer to the apache documentation.

Can I add header/footer to my directory listing?

Yes. Apache honors the .htaccess files that you place in the file hierarchy of your projects (files under /home/groups/<project_name>/htdocs/). You can therefore implement header/footer using .htpasswd. Please refer to the apache documentation.

Can I use php on my project's webspace?

Yes. We have enabled php for project's webspace. Nevertheless, the primary reason php is installed on the webserver is to support the web interface of InriaForge. Therefore, php.ini and the php version installed are configured to match InriaForge's needs. If a security warning imposes changes on the configuration, we will act without notice to project admins. If the change is the result of a planned upgrade, project admins will receive advance notice so that they can plan for the upgrade.

Can you create a database backend for my project's webspace?

No. The database server on InriaForge has for only purpose to host InriaForge's data. If you need a database backend for your project's webspace, please use other ressources.

How to redirect to another existing site?

Change the home page (file index.php) hosted by InriaForge. Put the following HTML code between HEAD tags : <meta HTTP-EQUIV="REFRESH" content="0; url=http://myotherserver/myproject">

About the Wiki plugin

How to activate the Wiki plugin for my project

A Wiki plugin is available for projects hosted on INRIA Forge and it is based on MediaWiki.
By default this plugin is not activated at the creation of a project.

To activate this plugin, you need to do the following as project's admin :

In the Mediawiki plugin page, you will find links to pages explaining how to use this wiki (eg. the text coding rules).

How to change the MediaWiki top-left icon?

Go to your project page, then Administration > Tools > MediaWiki Plugin admin.
You can upload a new logo, just follow the instructions on that page.

How to get notification of edits on a MediaWiki page?

About the docs area

I can not submit new documentation in my docs area

This happens if you try to upload a large documentation file. Unfortunatly, the exact file size limit is unknown and depends on the content of the file itself. Furthermore, no error message is displayed to explain that the submission failed because the file is too big.

To workaround this problem, you can upload your documentation to your web area and then fill in the OR specify an outside URL where the file will be referenced : field in the documentation submission form to point to that uploaded file with something like:

This workaround suffers from the fact that the uploaded documentation file will be publicly accessible to all, even for private projects, because it will be served by the web server with no access control.

How can I delete a document file uploaded in the docs area?

To delete a document file uploaded in the docs area (see "Docs" tab):

If you see no "Admin" link, your role on the project does not allow you to delete a document. So contact a member of your project who is "Admin" to do that or to modify your role to be allowed to. For more information about the roles, see this FAQ entry.

How can a tracker/forum send mail to a mailing list?

You might want a tracker or a forum to send mails to a mailing list, for example so that any new bug submission is announced in the developper mailing list.

First you need to be a member whom role allows to administrate the tracker or the forum. Then you can access to the "Admin" page relative to the tracker, or forum. Once on the "Admin" page :

The mails coming from the tracker or the forum are set such as the sender is :

So you need to modify the configuration of the list such as it accepts mails coming from this kinder of sender. To do so, see the How can I make my mailing lists automatically accept mails from some domains? entry.

The list manager may complain that the message coming from a tracker or forum was not an explicit destination and may reject it. To solve this, you have to modifiy the mailing list configuration. The option you are looking for appears under Privacy Options/Recipient filters. You should set require_explicit_destination to no.

Can a git commit trigger a tracker update?

If you want to automatically update a tracker within a git commit, this is possible. You just have to put the [#tracker_id] in the commit subject (first line of the commit message). The tracker will be updated once your commit is pushed.

Note that for this feature to be activated, there are 2 prerequisites:

1) the "SCM hook" tool must be activated (*)

2) the "Commit tracker" must be activated in the post-receive triggers


(*) Warning:

How are the statistics of the front page computed?

The front page presents InriaForge Statistics to all the visitors. On top of that, on the default page of each project displays an Activity Percentile for the project. These statistics don't always seem right. Here is a list of reasons they might not show expected numbers.

  1. The number of hosted projects seems low: this is because only public project are counted here. Better stats are available from InriaForge's admins.
  2. There seems to be no correlation between the box displaying the number of downloads and the one displaying the most active projects. This is because the first box is cumulative, whereas the second only takes into account the previous week.
  3. To be a bit more precise, activity percentile is computed from
    1. Activity in the project's forums
    2. Activity on the project's task manager
    3. Activity on the project's bug, tech support and patch trackers
    4. Number of downloads
    5. Number of released files
    6. Activity in the SCM repository
    Beware that the percentile reflects only the rank of the project regarding its activity computed as explained above. If N projects have an activity (i.e. there are downloads, releases or activity in forums, etc.), then the "most active" project have an activity of 100%, the second project has (100-(1/N))%, the third has (100-(2/N))%, etc. In general project with rank M has an activity percentile of (100-(M-1)/N)%. Only public projects are displayed in the list active projects.

How to configure Eclipse to work with InriaForge?

To manage your ssh connection with Eclipse, see HowTo configure Eclipse.

How to name forums, mailing lists on InriaForge?

Because of predefined reserved names on InriaForge architecture, when naming your forums or mailing lists, please AVOID theses words :

How to change the license for a project?

Resources of a public project (mainly the code) are supposed to be distributed under a legally valid license.
"None" is not applicable choice in this case.

To change the project license, you need to (requires project admin privileges) :

How to connect Hudson (continuous integration server) to InriaForge?

How to notify my Hudson server of commits on my Subversion repository?

Just edit the post-commit hook

and add the following statement:
/usr/bin/wget -o /dev/null http://<hudson-url>/job/<job-name>/build
To access secure (https), some further options are required:
/usr/bin/wget -o/dev/null -username=<bot-name> -password=<bot-pwd> -no-check-certificate -auth-no-challenge https://<hudson-url>/job/<job-name>/build
Using https access, the password is unencrypted in post-commit hook that is readable by all. Don't forget to change permission (all -r all +x)


How to transmit parameters to my job?

For builds configured, the URL is slightly different:

Note that the address ends with buildWithParameters instead of build, and the separator \& between the parameters instead of the usual & in URLs.


How to avoid project building systematically, if commit doesn't concern my job?

It is possible to run a polling on SCM rather than forcing the build, so the project will be built only if significant changes to the workspace have occurred. For that you must replace the end of the usual URL ("build") by "polling". We therefore get:

In this case, it is not possible to pass parameters. This also implies that the option "Scan Tool version control" is checked in the configuration of the project. You can not define polling schedule, leaving the field blank, and only rely on notifications scripted to have a purely push.


Many jobs are impacted by my commits, do I list them all in the post-commit hook?

When multiple jobs are to be built after a commit, there is an alternative to the complete reporting of all jobs. In the Subversion plugin, there is a mechanism to notify a change in the repository without triggering the build or polling a specific job. Hudson then automatically determines the jobs impacted. This solution is simpler but also more accurate because Hudson is based on the path of files modified in the changeset to identify jobs that need to be built. To do this, add this to the post-commit hook:

   UUID=`svnlook uuid $REPOS`
   /usr/bin/wget \
     --header "Content-Type:text/plain;charset=UTF-8" \
     --post-data "`svnlook changed --revision $REV $REPOS`" \
     --output-document "-" \

Warning: This strangely not build the revision passed in parameter in the URL, but the latest version (HEAD).

How to build the revision that corresponds to the notification (and thus build absolutely all revisions)?

It is possible to force Hudson to build the revision that triggered the hook, but the job requires a setup that goes beyond just tick a box as is the case for other methods of notification. It is necessary to create a job and set to pass the revision number as parameter. It then uses two things:

We must therefore define a parameter we call such REVISION. By default, it will be empty, it will mean "HEAD", otherwise it will contain @ddddd.

We write the URL of the repository as well:<project>/<path>${REVISION}

Through the mechanism of substitutes results, we will have a URL understood by reference to the plugin and the good review:<project>/<path>@dddd 

Then, you should define the field "local directory module" and repeat the last file path (this avoids the issue of our workspace is suffixed with "${REVISION}" not substituted ! )

Finally, it should trigger a build set from the post-commit hook (see the section devoted to this subject). The URL should look like this:

Note the 40% (char @) don't forget it, if not present the URL of the repository be incorrect.

Leaving REVISION undefined, the variable ${REVISION} will be replace with a blank (default value set in the project), and the URL of the repository will not be suffixed with @dddd and so Subversion will grab the last revision date.

I get the error 405: Method Not Allowed

Do not forget to give permission Job/Build for the user running the build in the security matrix (defined in https://<hudson-url>/configure ).
This can also happen if you try to pass parameters to build using the wrong URL. See the section on parameterized builds.

Miscellaneous other questions or issues

My commits are attributed to somebody else in the "Repository History" zone

This is because you committed/pushed to this project with another username or e-mail address than your Inria Forge account identifier. The algorithm displaying the "Repository History" tries to match this unknown identifier with a registered Inria Forge user, and may fail.
The solution is to put a .mailmap file at the repository root level (at the same location as the .gitignore file if you use such a file) to help the algorithm. More details about this mechanism may be obtained with "man git-shortlog".
For example, your .mailmap file may look like:

Firstnamename Surname <> <>
After creating this .mailmap file, you may need to ask the Forge team to recompute your project's history (please open at ticket).

About Spam and InriaForge

How can I make my mailing lists refuse postings from non-subscribers?

See this FAQ entry.

How can I prevent anonymous users from inserting spam in a tracker or a forum?

See this FAQ entry.

How can I delete a spammed (or unwanted) message from a forum?

To delete spammed message or unwanted message from a forum, first you need to be a member whom role allows to administrate the forum. If so, follow these steps:

In case of problems, mail the administrators or file a bug.