Le blog de pingou

To content | To menu | To search

Général

Entries feed

Tuesday, December 10 2013

RDFa with rdflib, python and cnucnu web

source.png

Fooling around with RDFa and some projects

Continue reading...

Monday, October 28 2013

F19 on jenkins

You may be aware that we have a jenkins server running in the fedora infrastructure for our projects (and some others).

Until today, it was running two builders: one EL6 and one Fedora 18.

I have now added a third builder running Fedora 19 and I need to look into adding a Fedora 20 once the beta is out this could really help for testing applications against the latest Fedora.

I will probably retire the Fedora 18 builder when I add the Fedora 20 one, so if you have a project building only on Fedora, you might want to check its configuration.



PS: If you have a project on jenkins, you may want to check if every thing it still set-up correctly, I just had to re-configure a couple of projects which are using git as scm.

Faitout - test against a real database

  • Do you do unit-tests?
  • Do you do continuous integration?
  • Do you use sqlite for your tests while deploying against postgresql?
  • Do you hate using sqlite for your tests?


If you answer 'yes' to any of those three questions, the following post is for you.

Otherwise, well, stay, it might still be interesting ;-)

When doing unit-tests you want to have something fast which allows you to quickly see if your last changes affect other part of your code.

sqlite is great for that. You can easily create in memory database, no FileIO, it all goes fast and smooth.

That is until you push your application to production where it is deployed against a real database system such as PostgreSQL. Then suddenly, queries which run fine under sqlite start breaking under PostgreSQL. sqlite and PostgreSQL implements some things differently and this leads to this kind of situation.

The solution for this is of course to run your tests in an environment as close as possible from the production on, ie: run your tests on the same database system as the one you use on production.

But this can also become complex, it means setting up a new database server, create a new database, clean the database after the tests, handle permissions...

With this in mind, project such as postgression appeared.

The idea is simple: easily get access to postgresql databases which are thrown away after a certain time.

The problem is that postgression is not FOSS, thus when a couple of weeks ago there was no way to get a database, there was also no way to set up our own postgression server that could be used by a restricted number of person.

So after discussing it with Aurélien, somewhere between lunch and dessert, faitout appeared.

The idea was simple, have a small web application, create on the fly a user and a database made available to the on who asks and after 30 minutes (via a cron job for the moment) destroy the database and the user.

The API is pretty simple and all is documented on the front page of the application.

So feel free to have a look at it, test it, break it (but let us know how you did that ;-)) at the test instance we have:

http://209.132.184.152/faitout/

Thursday, October 24 2013

Back on Nuancier

It has been a while but it has been about as long that I wanted to write this blog post.

I would like to come back a little on Nuancier.

Nuancier was born during the last flock in Charleston. The idea was to create an application to simplify the process of submitting and voting on supplementary wallpapers.

Until now the process was rather complex and not really user friendly.

  • People upload their artwork onto the fedoraproject wiki
  • Artwork would be validated by the administrators of the process
  • From there we would set up an election in our classic election application
  • The community would vote on names (the current election application does not have support for voting on artwork) which meant, having the list of submissions open on another tab/browser and keep navigating from one to the other

Not quite a user friendly process.

Nuancier aims at solving this. However, between flock and the last supplementary wallpaper election time was short, thus it was decided that if we could not manage to have a full feature app in place in time, we could manage to have the voting part done and deployed.

Thus we created nuancier-lite which is deployed and was used for the last election.

Among the features:

  • Single page to vote where all the candidates are shown and you can just pick the one you like
  • After discussions on the devel mailing list, you may claim a badge for voting
  • Integration with fedmsg (election start/stop, results are published)

What I would like to do is come back on the results of these elections.

Anyone that had signed the CLA and was is more than one group could participate to this election.

The statistics of this election show that

  • 128 person participated
  • in average people voted on 12 wallpapers (16 was the maximum number of vote allowed)
  • 52% of the participants voted on 16 wallpapers
  • 68% of the participants voted on 10 or more wallpapers

And looking at the nuancier badge

  • 79 person claimed the badge (61% of the people who voted)

Seeing that there are more than 3000 person registered on badges, I must say I was expecting a higher participation but according to the people used to run this election it is already much higher than it used to. On the other hand, we probably missed some annoucements as it was not even announced on the design-team list, something we'll have to be careful to do next time.

But for nuancier, there is still some work to be done:

  • Add the capability for artists to upload their artwork via nuancier
  • Possibility to name and pick a valid FOSS license for each of their artworks they submit
  • Automatically check that the import is valid (type of file, size of the artwork...)
  • Add the possibility for admin to moderate/approve submission using nuancier directly
  • Interact with the design team to make the whole process as user friendly as possible

It is even a project for the Outreach Program for Woman 2013, so if you are interested, jump in!

Thursday, September 12 2013

Fedora vote history

There has been some discussion recently about elections in Fedora and that gave me the desire to have a look at the history of our elections with regards to the number of participants.

I went back up to about 2008 and here are the results:

        Board   FESCo   FamSCo
			
2008-07   250    150	
2008-12   227    169      126
2009-06   297    308	
2009-12   225    216      130
2010-05   229    180	
2010-11   239    240      125
2011-06   204    200	
2011-12   224    225      104
2012-06   199    236      207
2012-12   201    206      102
2013-06   157    166      175

While I was digging the archives (I have the email announcement for each of these numbers), I also ran into the Fedora release name election results, so here are the evolution of the number of participants:

Release Ballots	
F9	62
F10	390
F11	310
F12	393
F13	313
F14	206
F15	296
F16	421
F17	292
F18	429
F19	391
F20	361

Wednesday, August 14 2013

Back on Flock 2013

Flock 2013 !

DSC_0958.JPG

Since my last blog post about my arrival to Charleston (South Carolina), I have been participating to the Flock conference.

This time, around 200 contributors from 19 different countries gathered to discuss, debate, exchange and more importantly work on Fedora.

This is briefly what I have been doing over these four days:

Day 1

After helping out with giving away goodies, badges, lunch cards and T-shirts, I assisted to the fedmsg presentation by Ralph Bean. It was a nice overview of the project as it stands today and also a good presentation of the environment that is growing up around it (datagrepper, fedora-mobile, Badges and many more).

Then I had to decide between HyperKitty and the talk "Why Fedora sucks", knowing Aurélien and knowing that I am able to speak frequently with him, I went for the later. Christoph explained that he went back on some criticisms made few years ago about the project and going through this list, a number of them have improved or even some solved. On the other side he also listed a number of current issue, on UI, on comps and other aspect of the project. So there is still some work for us and we can't all retire yet ;-)

DSC_0085.JPG

The last talk of the morning was from Haikel Guémar on FOSS development and Agile methods. I learned some things bout Agile itself and found out that some ideas I like are actually part of known project management processes. Quite cool all in all :)

The afternoon was dedicated to one larger discussion and a hackfest.

The discussion was about the future of IT in general and trying to get some ideas of where we would like to be in five years as a project. Quite an interesting debate started from this with ideas bouncing around of where people believe we will be in five years. Time will say if we were right during this session :)

The second part of the afternoon has been dedicated to a hackfest on semantic web technologies and more specifically how we could include the ontologies in our application. Using the DOAP ontology we can provide a standard way of describing a project, using the same standard as launchpad, pypi or debian. This should make data integration that much easier in the future.

For the evening program we had dinner in a restaurant few minutes walking from the hotel. We had a good time, it was hot and quite humid but we were outside and the fans helped us quite a bit.

Day 2

DSC_0957.JPG

The day started with an awesome talk on 3D printer by Aeva Palecek from Lulzbot. Unfortunately, as pointed out by kevin earlier are out of stock for their printers :-D.

After that I went to the presentation of Suzanne Reed from FOSSBOX RIT the project she worked on gourciferous based on gource to visualize the history and evolution of git-based project. Quite a nice tool and a nice presentation as well.

Following this presentation, I went on learning a little more about darkserver and what it can be used for. Kushal had already introduce it to me some time ago but I wasn't sure I had understood it completely, while I have a much clearer idea now :)

DSC_0989.JPG

For the last talk of the morning I went on learning a little more about clouds and Eucalyptus. Greg DeKoenigsberg is a really good speaker and the presentation (even was unfinished slides) was really really nice! It gave me a clearer representation of the cloud stacks available these days and how they differ form each other. Really interesting, thanks Greg!

The whole afternoon has been spent with the infrastructure to figure out how we would design and eventually setup a AuthZ server using some of the 0Auth principles (but not all).
The idea is to support the use case of people running job via CLI or cron against the Fedora Infrastructure but limiting their range of actions by providing them certain tokens that restrict them to the given action. (Am I clear here?) This is something we want to work on, so keep in touch if you are interesting, there are more thing coming ;-)

In the evening, we went to a bar, pretty much right in front of the hotel. The food was nice and the atmosphere really to 'relax'. We ate, we drunk, we danced and we had a very nice time over there. Toshio and Aurélien are making such an amazing dancing duo! Do come to Flock, if only for that ;-)

Day 3

DSC_0013.JPG

The keynote speaker of the day was the author of the Cantarel font, used by default on Gnome 3. The talk has been about font development, open-font, he touched upon the web-fonts as well. It was quite interesting, less in my usual area of interest so most things were new to me which was nice.

I was supposed to give a presentation after that session but when I showed up in the room, no-one was there. I waited until ten past the hour and nobody showed up so I just went to Ricky's talk on code review for Fedora apps. Ricky did a nice job in presenting the advantages and limitations of code review as well as listing the different points to pay attention to when doing one.

The following talk had been by our local OpenID expert, Patrick. He presented us with how OpenID works in general and what are the extension that he developed for the Fedora Infrastructure, to satisfy our needs. All our application are slowly moving to OpenID, providing a single, central place where the user gives his password. In theory, ultimately, our contributors should not have to enter their password in any other place than the OpenID server.

And the last presentation of Flock I went to, was the presentation about Census by Nathaniel. Census is the replacement for smolt. It's still work in progress but the progress made are really nice. Nathaniel presented us with the design of the application, I must say I really liked its simplicity. I do think it should scale well and we should be able to make something of it. The bonus is that on the contrary to smolt, Census is design in such a way that we can throw any kind to data at it. So it is another place for us to collect and provide statistics about the project and its contributors.

The afternoon has been spent hacking on fedocal with Haikel. The next release is almost out of the door. Haikel, lbrabec and I were able to close the last remaining tickets for the version 0.2.0.

For the last evening, the organizer planned, I think, one of the most awesome dinner location I can think off. We had the full aquarium of Charleston for ourselves! From 7pm to about 10:30pm, we had drinks and food at the aquarium, looking at fishes, some snakes, an impressive white alligator and a very cute bird that spent 10 minutes seducing tatíca! It was an amazing evening a big thanks to the organizers for this awesome opportunity!

Day 4

DSC_0069.JPG

The last day was dedicated to hackfest. So I hanged around with the infrastructure team. We were able to make a list of the coming tasks we want to do as well as those that we want to spend more time discussing about. After lunch, when we all moved to our tasks, I spent some time testing the new fedocal, fixing some of its configuration files and adjusting its spec file, testing the DB upgrade, only a couple more things to check/do (such as the update of the documentation) and I should be able to push this new version.

DSC_0092.JPG



I have had a really good time with everyone in Charleston. I met new people, made new friend, got work done and work planned and we just had some good times all together.

The Fedora project is full of awesome people and every time I meet them I realize how lucky I am to be and work with them :)

A big thanks to the organizers and Fedora for making this event possible and giving me the opportunity to be part of it.

DSC_0956.JPG Organizers at work ;-)

Wednesday, August 7 2013

Arrived in Charleston

Yesterday evening, after about 18h of traveling I landed in Charleston where will be held from the 9th to the 12th Flock.

So this morning I was able to walk through Charleston and discover this nice city, here are some of the pictures I took:

DSC_0748.JPG

DSC_0782.JPG

DSC_0785.JPG

DSC_0797.JPG

Some more are on my gallery



Images under CC-BY license.

Wednesday, July 10 2013

Small tribute

Just added this to my blog, it will remain there for some time, mourning such a loss will need it.


If you are interested:

<a href="https://lists.fedoraproject.org/pipermail/announce/2013-July/003174.html">
  <img style="position: absolute; top: 0; right: 0; border: 0;"
       src="http://blog.pingoured.fr/public/rip_seth.png" alt="RIP Seth" />
</a>

Tuesday, May 14 2013

Talks / Hackfest @ Flock

1 talk and 3 hackfests:

  • Fedora vs the semantic web
  • pkgdb2
  • fedocal
  • DOAP in fedora-packages

I'm not sure if any will be selected but we shall see :-)

Thursday, March 28 2013

GNOME-tagger

Surprise!!

Over the last few weeks Ralph Been and I have been working on the new version of tagger.

One of the idea of this new version is the integration with gnome-software that Richard Hughes introduced at the beginning of the month.

In order to do so, this new version comes with a clearly defined API.

To be completely honest, Ralph and I had already started to think about defining an API before Richard's post on gnome-software. One reason for which we started on this is for gnome-tagger:

tagger.py3.png

GNOME-tagger is a desktop application, writen in python and GTK3, developed as a GNOME application which is an alternative to the tagger web application. With it you can, add tags to packages, vote on tags (up vote or down vote), see the statistics about tags on the fedora package collection as well as see who is winning the tagger game. Pretty much all what you can do from the web application, but locally :-)

tagger_menu.png

This is still a little bit work in progress as we still need to assess how we want to handle authentification and the blacklist of tags as well as anonymous tags but hopefully we will have this new version ready soon and you can start playing with gnome-tagger!



Note: The new tagger API supports rating apps as well, but at the moment this something integrated with gnome-software but not with tagger web application or gnome-tagger.

Monday, March 18 2013

Fedora-Infra: Did you know? -- pkgdb-cli

Did you know?

With pkgdb-cli you can give someone commit rights without her/him asking for it?

The command will look like:

pkgdb-cli update <package> <acl> <user> <branch> --approve

For example:

pkgdb-cli update packagedb-cli commit toshio devel --approve


Can be handy if you work in a team!

Monday, March 11 2013

Fedora-Infra: Did you know? -- copr-cli

Did you know?

You know copr right ? This tool to create easily your own yum repo building the RPMs in a new cloud instance everytime. It allows you to give access to nightly builds or a new version of a software backported from rawhide.

Well, if you have been playing with the development instance, know that in the sources there is a CLI tool!

Some example of what it can do if run directly from the sources:

$ python copr_cli/main.py -h
usage: copr-cli [-h] [--version] {create,list,build} ...

optional arguments:
  -h, --help           show this help message and exit
  --version            show program's version number and exit

actions:
  {create,list,build}
    list               List all the copr of the provided
    create             Create a new copr
    build              Build packages to a specified copr

And a small demo:

$ python copr_cli/main.py list toshio
name                |description                                                                                           |repos |instructions
-----------------------------------------------------------------------------------------------------------------------------------------------
fas2                |RPMS for the Fedora Account System run in Fedora Infrastructure.                                      |      |
packagedb           |Package Database that manages ownership of packages for Fedora.  Includes both the client and server. |      |
flowhub             |flowhub adapts the gitflow workflow to github repositories                                            |      |
python-fedora-devel |                                                                                                      |      |

Some documentation on how to setup copr-cli is present in the README in the source

Monday, March 4 2013

Fedora-Infra: Did you know? -- apps.fedoraprojects.org

Did you know?

You cannot remember what's this application you've seen on the Fedora infrastructure at the latest FUDCon?

You cannot find the URL of that specific application of the Fedora infrastructure?

You are looking for an overview the applications run by the Fedora infrastructure?


apps.fedoraproject.org is there for you.


apps.png


Developped by Ralph Bean apps.fedoraproject.org gives you an overview, introduction and links to all the application run by the Fedora infrastructure.

Monday, February 25 2013

Fedora-Infra: Did you know? -- HyperKitty sends email

Did you know?

Do you know about HyperKitty? It is this new interface for mailing-list archives offering a forum-like experience for those that prefer forum over mailing-list.

The latest version (which you can see online on the development server) allows you to

  • create a new thread in the list as you would create a new thread in a forum
  • reply to an email sent on the list as you would reply to a post in a forum
  • tag posts on the list
  • like/dislike post sent on the list
  • get an overview of the list activity (see for the Fedora-devel mailing list)


hk.png


Want more info?

Checkout:

Thursday, February 21 2013

Subsurface 3.0

SubsurfaceBanner20121.png

Divers head-up!

I have just update Subsurface to its latest version on Fedora 17 and Fedora 18, it brings a whole bunch of new and shiny features:

  • GPS / Map integration to place your dives on the map
  • More dive computers supported
  • Decompression ceiling calculation using the Bühlmann algorithm
  • Dive planner
  • Support for multiple dive-computer (as in multiple dive computer for a single user)

And much more...


Feel free to test them:

Sunday, February 17 2013

New board for Borsalinux-fr

Borsalinux-fr (the french NPO dedicated to promote Fedora in the french speaking countries) held its annual general assembly this afternoon.

During this assembly we reviewed the activity of the NPO last year as well as the state of its account.

All in all, 2012 was a reasonnbly good year, the usual activities were held all went fine and there was the success of the FUDCon in Paris as a highlight. From a financial point of view, things are pretty stable, the budget looses a bit a money overall in 2012 but that's due to some reimbursement that happened in 2013 (and thus out of the overview).

Finally, there was the election of the new board, elected for two years. The board for 2013, 2014 is therefore:

  • Emmanuel Seyman (eseyman) - President
  • Pierre-Yves Chibon (pingou) - Vice-Président
  • Nicolas Chauvet (kwizart) - Treasurer
  • Pablo Martin-Gomez (bouska) - Vice-treasurer
  • Charles-Antoine Couret (renault) - Secretary
  • Kévin Raymond (shaiton) - Vice-secretary
  • Guillaume Kulakowski (llaumgui)


I think this is a nice board for the coming two years and I already look forward to work with all you guys.

Wednesday, January 23 2013

Back on FUDCon Lawrence

English version

Last week-end, in Lawrence, Kansas, the Fedora community held its North American FUDCon (Fedora User and Developers Conference).

I have had the chance to be able to go there.

I was great!!The organizing committee (including, Ian , Ruth and our dear FPL Robyn) did a grea job! A big big thanks to them.

On the friday I was able to:

  • Participate to the infrastructure session in the morning when we presented what we did over the last year and some of our plans
  • Attend the talk on fedup (the new upgrade tool from Will Woods
  • Give a lightning talk on fedocal, a web-based calendar application for Fedora
  • Give a presentation about micro-web framework and designing maintainable web application. I must say the audience was pretty small, but I think there were interested, maybe even convinced at the end. Yup, all three of them :-)
  • Attend the talk on Ansible
  • Attend the end of the talk on copr
  • Attend the talk on "Saving spins and Fedora Formulas"

At the end of the day, we had pizza in one of the big room of the hotel and I was able to discuss about future plans with different people from the infrastructure team, that was really a nice evening.

I spent the Saturday and the Sunday with the infrastructure team again, we have hacked on a number of things as reported by Kevin or Ralph.

Personnally, I have been working with Toshio on packagedb. We remove some of the older code that is no longer maintained and has been ported to our brand new packages application. This change will go live in ~2 weeks.

On the Sunday, I worked a little bit on copr trying to get the first steps into making an API and a CLI for it. It is still work in progress but some patches went through :-) I also tried to give a hand to Patrick who has been busy, although being in Europ, moving the OpenID server that we run within the project into its own container.

At the end, I think for me, what I enjoyed the most has been the meet all the guys from the infrastructure team with who I have been in contact for almost 18 months. I finally got to meet them and that was very nice.


So, Toshio, Seth, Kevin, Luke, Ralph, Aurélien, Tom, Ian, Ricky, Matt, Xavier, (and all the other that I have met and who participate in this great atmosphere of FUDCon) thanks a lot for the great time I've had, and I hope to see you again soon.

Like Ralph said, The Fedora community is full of some pretty awesome people.


Finally, I would like to thank the Fedora project and Bluehost for sponsoring my travel and thus allowing me to participate.

Friday, January 11 2013

Announcing fedocal 0.1.0 alpha

fedocal is a web-based calendar application for Fedora

Continue reading...

Tuesday, December 4 2012

python-pandas vs R

Speed comparison of python-pandas and R for some spearman correlations

Continue reading...

Monday, October 22 2012

Back on FUDCon Paris (2)

I have already presented what I have been busy with on the first day of the FUDCon EMEA 2012, now is time to present what happened on the other days.

Second day: Barcamp

On the second day we had the barcamp.

I attended a number of them


The Fedora User Experience talk from spot, great talk on how we should/could improve user experience in Fedora. There are a lot of ideas in the air and I am really looking forward to see them coming to life!


The I gave the barcam Cloud @ Infra, let's keep this one for now, I'll come back on it just below.


Then I followed the presentation from Kévin on the Fedora websites. He is doing a great job with it even if it seems to be messy sometime. If you're looking for a nice place to help Fedora, feel free to contact him!


Right after this talk, I have been able to see the end of spot and Ruth about the Raspberry Pi. I got one not so long ago and I am still starting to play with it but it is definitevely something fun to do.


I followed then the Kernel talk from Josh, very interesting even for someone like me who never really had to deal with the kernel (except for the traditionnal Graphic/Wifi bug). Josh already presented his talk and the consequences on his blog so I won't come back on it. Although I still have to ask him where I can get rawhide non-debug-activated kernel to test on my machine at work which gives me weird graphic glitches.


The last session of the day was with the ambassadors team, where we discussed a couple of subjects including mentors in EMEA and how to "filter" the list of ambassadors on the wiki to remove from it the ones which are now longer really active and that new comers might still try to contact without getting answers. One solution might be to simply check, while generating this list, if they have logged in using their FAS within the last 6, 9, 12 months or so. That would prevent from removing people from the group while still keeping a reasonnably up to date list of active ambassadors on the wiki (and bonus point: it easily allows people to come back!). That was one of the solution proposed but I am sure there will be some more discussions before the group agrees/settle down on one.


So that's pretty much it for the Sunday. Quite a busy day but also really really interesting!


Now let's go back on the barcamp I gave

Cloud @ Infra

So the idea of this talk was to see if we could come up with ideas on how we could use the cloud that we recently got in the Fedora infrastructure.

And we came up with some ideas :

barcam_cloud_at_infra

  • The first one is an idea I have been working on for a couple of weeks now, use the cloud to run a Jenkins instance with build nodes on differents OS (Fedora or EL). Eventually we could even offer this as a service for project hosted on fedorahosted.org.
  • One other idea would be to use the cloud to be able to use a simple build environment giving developers/packagers a similar tool as the PPA on Ubuntu.
  • Of course the cloud could be use to provide testing VMs on different distros
  • The cloud could be use to build personnalize VMs images, and maybe with a nice GUI tool on the top
  • The QA guys could use the cloud for a testing environment
  • Finally, one idea which I really liked: using the cloud to run automatic/periodic FTBFS. We could use the cloud to rebuild every packages that at branching have not been built since the last branching. Some were even proposing, rebuild every three months, all the packages that have not been rebuilt in the mean time. I think this would be pretty nice and woul help tracking down packages that fail to build on a newer Fedora for X, Y or Z reason.


Third day: Hackfest

The last day of this FUDCon was dedicated to Hackfest. A number of them have been proposed,

hackfest_fudcon_paris

I have been involved in:

And I also worked on a webapp, which I should present more thoroughly in a near future (suspens...).


This is the end of my report for the FUDCon 2012 EMEA in Paris. I had a great time and I would like to thank all of those who had the hard task to organize this event, especially Kévin who had the lead and did a large part of the work.


Next FUDCon, Lawrence, Kansas, US in January. I am already looking forward.

- page 4 of 10 -