Archive for the ‘FLOSS’ Category

Meet at Open Source Experience Paris


After a long time staying quiet, I’ll be again talking at an Open Source Event ! The Open Source Experience Paris is an old new event (having been called POSS, cancelled last year, or Solution Linux even before) during which I’ll talk about two topics in french this time:

  • 09 november 2021 11:20 – 11:40 – Room 241 – API != REST – procmail à la rescousse, or how an modern API can also be done using old tools such as the SMTP protocol and procmail, and also more “standard” REST APIs as well, all combined to help build a Workshop on Demand service. And I’m happy to present this topic with the help of my HPE colleague Frédéric Passeron.
  • 09 november 2021 14:20 – 14:40 – Room 242A – Vous souhaitez passer votre projet en open source ? or how putting your code developed with love during 2 years on is probably the wrong thing to do to get all the benefits of Open Sourcing.

I’ll be around the rest of the event, probably not too far from the Mageia booth as they have been kind enough to make some advertisement for my presentations !

It’ll be great to be able to see some of you as it’s a pretty long time since we were able to do so and discuss about the latest news in our communities. So see you there !

FLOSSCon 2022 ?


L’association FLOSSITA qui a organisé en 2018 et 2019 l’événement FLOSSCon se propose de sortir de son hibernation forcée, et d’organiser à nouveau une FLOSSCon en 2022 sur l’agglomération grenobloise.

Pour cela, il faut donner à l’association les moyens de le faire correctement : 

  • Des moyens humains avec un noyau de 5/6 personnes (un bureau) qui œuvreront à l’organisation (logistique, comm, informatique,…),
  • Des moyens financiers (cotisations et dons), d’autant que nous maintenons notre volonté d’avoir un événement gratuit, et qui permettent de couvrir les frais informatiques, assurance, communication, sans que tout ne retombe que sur les 5/6 bénévoles organisateurs,
  • Un lieu de conférence adapté
  • Des conférenciers essentiellement locaux désireux d’échanger sur leurs outils libres préférés,
  • Un public qui vienne écouter les sessions proposées.

De façon à mesurer l’envie de chacun de voir se concrétiser cette FLOSSCon 2022, nous vous demandons de remplir un petit sondage de 3 questions qui nous permettra de mesurer l’engouement (ou non) pour un tel événement.

Dans le cas favorable, nous réunirons prochainement les 5/6 personnes volontaires pour travailler sur un retro-planning et commencer à planifier la FLOSSCon 2022. Dans le cas contraire, nous ne renouvellerons pas nom de domaine, hébergement, et nous contenterons d’être conférenciers dans les événements des autres. 😉

Article sur le Logiciel Libre chez Combodo


Heureux de voir publié un article que j’ai commis sur le Logiciel Libre (FLOSS) pour un partenaire de longue date d’HPE et un acteur majeur du libre en France, grenoblois qui plus est 😉

Et regardez de près iTop, c’est une solution vraiment intelligente pour la gestion de ressources et de services.

New Mageia Board


After releasing its 8th version, the community members part of the french 1901 law association had a meeting tonight to reinvigorate the work done by the association in support of the distribution, and elected a new board to support that goal.

I’m proud to have gathered the trust of my fellow association members and been elected as a board member ! We also have a brand new chairman (neoclust, more than 2000 packages maintained !), a new secretary (auroud85, mister QA) and a new treasurer (maat, French users community).

With that new team in place and acting, I’m sure you’ll hear more and more about our distribution through various channels !

Mageia 8 is out


As announced on the project site, Mageia 8 is now out ! It’s great to see the community succeeding in another joint work to publish that one.

Now, I have to update my servers, my laptops, my workstations in order to benefit from it. Won’t happen tomorrow however, as each time I need to check for major release updates and potential migration aspects before doing so. Using it for all what I do from apache, freedb, squid, shorewall to gnucash, tellico, digikam, and of course the usual firefox, thunderbird/mutt, libreoffice, that’s quite some software to test. Not counting the one I need to use on top of it such as teams, hpemyroom, zoom, openconnect with a Yubikey for my work.

Will start to look at shorewall first and fix my ansible playbooks to deploy the conf once the machine has been updated. Some fun ahead 😉

Congrats to all of you who contributed to that release (myself included in fact !!)

Our video is now available


It was great to be able to speak during again this year, even if being down under would have been even greater 🙂

The video of this great conf (!) has now been published so if you have missed us (Frédéric Passeron and myself) speaking about the Wokshops on demand live, you now can view it online. Enjoy !

Speaking at 2021


This conference is special for me. It has been one of the first big FLOSS conference WW to which I’ve been accepted as a speaker. It was in Sidney in 2007 and that was also an opportunity to organize the first and unique WW MondoRescue meeting (we were 2 !!). And I keep a strong feeling about that presentation and the australian way of life I really appreciated a lot.

Then I was luck to be selected again two more times in 2013 and 2014 in Canberra and Perth for a cross-distribution miniconf I organized there and a talk on Again great travels and events so well organized and allowing to meet all the greatest FLOSS contributors. Just a dream !

And that’s there again. Well quite ! I’ll be speaking again during the 2021 sysadmin mini-conference with my colleague and friend Frédéric Passeron around the work we’ve done to make Workshops on demand accessible to everybody. This session will cover all the gory details behind this project made of ansible, scripts, procmail, REST and SMTP API, and more !! in particular explaining why an API, even in 2021 is not necessarily a REST one to be useful 😉 All that in less than 20 minutes.

Watch out the agenda to compute the correct time for it in your timezone (yes it will be late for the 2 of us based in France), but that’s a fair price to pay to be honored to contribute to one of the most interesting conference on FLOSS (with Fosdem !).

One regret ? Well, yes, of course, not being in sunny Australia with Frédéric to enjoy what happens outside the conference itself !! But as much as I dislike doing virtual conferences, this one is so special for me it deserved efforts on my side. See you there hopfully !

Announcing the GeneanetForGramps plugin


I’m a happy user of Gramps, since the confinement started in France back
in March 2020. I had received since the death of my father a lot of documents that described my family, and I had also some holes in my understanding of some branches. I wanted to keep all these information easily on my computer, instead of on multiple unrelated sheets of paper. After looking at what was available, I decided to invest more time on Gramps and it turned out it was a perfect fit for my needs (well in fact it should be perfect for a lot of amator genealogists considering the high number of features it provides): ability to keep family trees and visualize them in multiple nice ways, ability to also store people images, as well as official acts related to their life, their profession, address and much more again.

The nice stuff for me was that Gramps is written in python, so could help me enrich my understanding of the language I’m practicing since 3 years now (even if I still prefer perl ;-))

And when you are looking for information around family members, you often find good details on Geneanet. It allows the sharing of numerous family trees published by many people, among them you’ll find some common ancestors ! And once you’ve found one, you can see the full tree going up, and you start copying information from it, to enrich your own tree. Of course *real* genealogists do not do that as they want to check with acts that this is indeed true, but I just have one life, and as with FLOSS, I like to rely on work done by others and just do new stuff to improve the global knowledge.

But it’s soon very painful to copy from Geneanet data that have been entered by others, because it takes time, it’s error prone, and a machine could do that instead no ? Well, Gramps has no way to do it, nor did I find anything related to exporting data from Geneanet. So python to the rescue 😉

I started by writing a script to extract data from Geneanet, using the lxml module which is very helpful to detect the tags that are in the Geneanet page and thus gather the information I needed to store that in a python object. Then, using the Gramps python module you can find whether the person already exists in the database and either include him or her if not, or modify (or not) the attributes if he or she was there. Then you add recursivity for ancestors to parse the tree, then you add recursivity for descendants to complete the tree, and then you add spouses support to cover it all. It took me a bit of time to make that work, but by june I had a good working setup. I was able to import in my test DB lots of people in one run, from the command line interface. Not everything was done, but as we were free tomove again, I used more to time to attend shows, exhibitions, visit museum, enjoy concerts again…

But as we were confined again in october, it was time to start thinking about the graphical interface, because at the end, I want to make it easy for my wife (and myself) to make such imports. Took me a bit of time to figure out which example was the closest to my case, in order to get inspiration from it, especially for the various parameters I wanted to make available. And this is now done and working as shown with this screenshot:

GeneanetForGramps plugin interface

Up to that point, I was still using the french web pages for Geneanet, but of course, if you want to make your plugin used world wide, you have to deal with internationalization. So that was the last step, replacing all hardcoded chains in the program by call to gettext and to translate stuff. Of course, it’s a bit mre difficult when you do string detection in pages, but it turned out that the code was clean enough to make it painless. And voilà ! I now have a working plugin that I try to integrate to the upstream project (see my pull requests). Hopefully it will be seen as useful and validated soon. Meanwhile you can integrate it yourself in your Gramps instance by downloading it from my Github repository as it’s AGPLv3.

project-builder 0.16.1 is now available


When you start updating something, you often finish working on a unrelated but necessary topic before being able to finaly solve your first problem 🙂

So I had received remarks on issues to install MondoRescue on latest Ubuntu vesions (18.04 and 20.04) because of the lack of afio and also star in the distribution. Indeed both have been droped in Debian since quite some times in fact 😦

So in that case you have various choices :

  1. give up (not my state of mind anyway !).
  2. point them to the sources of the tools so they help themselves.
  3. move to more supported tools (that’s something that is now planned).
  4. provide users with the appropriate packages.

The two last possibilities are the nicest to me and the users. So I first build a round of packages for afio for all latest Debian and Ubuntu distributions, at least to unblock users. I already had the build infra in place, and a spec file for rpm, so wasn’t too hard to do it. Next one on the list is s-tar. And that’s where the perturbation happened 😉

I realized I had a hole in project-builder as it was always supposing that upstream tar files were in tar.gz format. That’s not a problem when you develop the upstream project yourself, because you generate indeed these files, but when upstream is s-tar and provides tar.bz2 files, suddenly you have a bug to solve 😉 Interestingly enough I didn’t had any issue since 13 years I’m working with the tool.

It wasn’t that hard to fix as the places referencing a source tar.gz file were easy to find, but that had implication also in the 150+ VMs and CTNs I’m managing to build packages. All of them also needed to have a update in order to support that correctly. Now project-builder doesn’t make any assumption on the nature of the tar file (but still requires a tar file BTW). A lazy solution would have been to just unpack/repack s-tar and voilà ! But I want to be able to use the upstream delivery untouched (checksum included).

So due to that 0.16.1 (to show incompatibility with 0.15) is on its way to the ftp server, and built at the moment for the various distributions. Then i’ll have to update all my VMs and CTNs with that new version, and finally I’ll be able to generate s-tar packages.

Then I’ll be back to what I’m really working on, which is the 3.3.0 of MondoRescue !!

HPE now contributes to OpenBMC !


For those of you reading articles on this blog, this is not completely a new info. Indeed, during Fosdem, it was already possible to see a demo unit of a HPE ProLiant DL 360 running Linuxboot and an iLO 5 running an OpenBMC software stack.

But that was a proof of concept. Now, HPE has officially announced on the OpenBMC mailing list that we were contributing to the project and will bring our patches upstream, and help the community build a better software stack. I supppose that in particular our knowledge of the Redfish standard will be helpful to improve the project.

And more over, if your company is interested to test such a combination, feel free to contact Jean-Marie Verdun, Mike Garrett (both writers in the thread and promoters of this internally) or myself and we should be able to provide to you an access to such a system.

But that’s not all, we’re also…. hum, well you’ll know when it’s offically announced 🙂

Anyway, that’s the best news for our Open Source involvement since monthes and I’m glad that we’re making that possibility a reality for all the customers we have wanting a more open platform.

Stay tuned, more to come !