Cazou's pastebin

GSoC 2011 - Week 4

2011-06-20T03:05:00.000-07:00

Hello everyone !

This week, I almost finished working on the portage_2_2 backend's access functions.

Thos functions are used to get information on packages and on the environment from the plugin (in this case, Portage 2.2).

There is an exception for get_property() which currently loads up all properties for all packages and returns the one asked for.

The next step for this backend is to implement actions : install, uninstall and update a package.

The implementation of those actions in the portage API being not ready yet, I've also started working on the pkgcore plugin.

Here, I have some troubles as to some functionalities are not present in pkgcore. Or at least, not accessible. To name a few, versions comparison, hard masking reason, package size. And I certainly haven't got into other ones. So the question is : do I use portage API's functions to do this or do I reimplement those ?

Another possiblity could be to have functionalities not implemented in a backend and make porthole aware of it. Messages like "The Pkgcore backend does not support this feature" would be shown instead of the masking reason. Of course, functions like version comparison must be implemented.

The API being a lot different from portage's, I found some functions that could be simplifyed or even removed, because they where doing something similar.

As the project is reworking porthole (and not just interface qith a new backend system), there will be rework to remove and improve some functions too. Particularly, the best() method which compares a list of versions (with no package name). Pkgcore allows me to compare versions with package name. I don't see the point in keeping the function that way.

This week, I'll spend some time on studying for my last exam (Friday) and continue working on the pkgcore module.

GSoC 2011 - Week 3

2011-06-20T03:04:00.000-07:00

Hey !

This week, I've been very busy, expecially with an exam that has been moved because a professor forgot to come...

I've been able to finish and fix problems with the Option Chooser.

There is still a bug I don't really understand but that doesn't seem to affect the general application behavior.

This week, I'll fix and do some clean up in the lib.py functions of the backend. I'll also start working on the different emerge functions : install, remove and update a package.

GSoC 2011 - Week 2

2011-06-20T03:01:00.000-07:00

Hello everyone !

This is the second status report for my GSoC project, regarding porthole code

rework.

This week :

I've been fixing issues with the dynamic options chooser widget.
I've been making it use the proper options
I've been doing some testing of the port to the new backend and checking if everything has been properly ported (The get_property method must be reviewed)

Time is hard to find as I have exams now and later this month. The next one being on tuesday.

The next week,

I'll finish that option chooser (making it impossible to select incompatible options, show correct tooltips)
I'll finish checking and testing functions that have been ported to the public library

GSoC 2011 - Week 1

2011-05-30T13:00:00.001-07:00

Hello everyone !

So, I've started coding on my GSoC project. If you don't remember what it was, see this blog entry.

This week :

I reviewed how would the backend work : TerminalOperations won't be a backend like APIOperations or the future pkgcore backend. It is a class that will be used, depending in the user preferences, to make operations through the old Terminal api.
I made the backend load it's configuration from a file.
I also made it use the portage public API functions and classes instead of Porthole's
I made the new main window use only operations from the backend (currently portage_2_2)
I've started changing the Advance emerge dialog to dynamically load options from the backend config file

Next week, I will

Fix issues with the dynamic options chooser widget.
Make it use the proper options and show correct tooltips for each options from the config file
Do lots of testing the port to the new backend and check if everything has been properly ported (I know it's not the case yet)

Note that since the beginning of May, school is taking me an awful lot of time (first projects, now exams)

But the exams session being fairly light, I take some time now.

GSoC 2011 !

2011-04-29T03:48:00.000-07:00

Hello everyone !
This year, once again, I've been selected for a Google Summer of Code with Gentoo. I will once again work with portage and it's GUI's.
The goal of this GSoC is mainly to rework Porthole's code to use different Portage back-ends (the terminal back-end and the portage API back-end).

The terminal back-end is already almost done and must be stabilized. Then all the terminal back-end features will be implemented in the portage API back-end. That will give a good idea of what should be in that API.

Brian Dolbec (Porthole's maintainer) will be my mentor.

This is my first actual project in Python and especially in pyGTK.
I'm really looking forward to learn pyGTK and Portage's internals.

Each week, I'll post at least one ticket on this blog to show the progress of the project, what goes wrong, what goes right and what I learned.

GSoC - Week 7 : An API

2010-07-12T08:55:00.000-07:00

Hello there !
Week 7 finally paid off, we have a C api for layman, based on Brian Dolbec's new Python layman API.
It can be downloaded from my git repository :

git clone http://soc.dev.gentoo.org/~cazou/c-layman

To test it, you'll need Brian's Python lib. You can get it from git too :

git clone http://dev.gentooexperimental.org/~dol-sen/git/layman.git

and switch to the API branch

git checkout api

(if this last one does not work, try git checkout -b api remotes/origin/api)

As API's are not stable yet, it is possible that it sometimes won't work at all, Try to always use the latest git revision for both libraries.

The method I decided to use is the interpreter one. Calling a python function from C becomes as easy as :

int bareConfigSetOptionValue(BareConfig* cfg, const char* opt, const char* val)
{
PyObject *obj = PyObject_CallMethod(cfg->object, "set_option", "(zz)", opt, val);
if (obj)
return 1;
else
return 0;
}

which is quite handy.

The C lib is still crappy with memory management and does not have a building system yet.
To use it, enter the 'src' directory and run :

gcc -W -Wall -L/usr/lib/python2.6/ -lpython2.6 -I/usr/include/python2.6/ -o tester --std=c99 -g *.c

Using python2.6 is important here.
At runtime, you must let python know where the layman api is :

PYTHONPATH=/where/you/put/layman.git/:$PYTHONPATH ./tester

It should update the overlay list and print that list.

Next week, I'll work on getting it stable, fix all memory leaks, document it properly, even within the code and rename some function like interpreterInitialize() --> laymanInitialize()

GSoC - Week 6 : One problem, 2 solutions

2010-07-06T04:37:00.000-07:00

Hi there,
This week has been the one were I realized that most of the remaining bugs in Kuroo can't be solved without rewriting some parts of the emerge/portage part of the code.
Instead of doing that, I decided to focus on the C portage library, which will eventually solve those problems.
So, if you don't remember, a C portage API will be written, based on a new python API, so that non-Python applications can easily use portage.
There must be a way to run python code from C and interpret the results in C. And there are 2 ways of that :

Running a python script which uses the portage API and gives the result on STDOUT, then redirect this output to the library by a socket.

Directly embed Python in C with the C Python interface directly provided by Python.

The first solution is the one used in PackageKit. The reason is most likely because the portage API changes a lot and it is easier to modify a Python script than a C program using the C Python interface.
The goal of having a new portage Python API is to get to a stable API for portage. The stability of this API is a good reason for using the second option.
The second solution is also less demanding in CPU (no need to fork, use a socket, redirect output and parse it)

My guess is that we will use the C Python interface. I actually wrote some code to test all that and you can access it from my git repository :

git clone http://soc.dev.gentoo.org/~cazou/c-portage c-portage

There is no building tools yet but the compile command is at the start of the source files (which are 2 different programs).

There is also the kuroo4 live ebuild online. You can access it by :

git clone http://soc.dev.gentoo.org/~cazou/ebuilds ebuilds

This repository will also contain ebuilds for the C portage API.

GSoC - Week 5 : Is there someone ?

2010-06-29T12:06:00.000-07:00

I only worked 10 hours this week, being busy with lot of other things after the exams.
Fortunately, I managed to do some interesting stuff :
Fix progress bars
Fix freeze at startup.
Fix text filter in PortageTab
Fix filter group
Fix package inspector

It becomes hard to crash Kuroo now. At least, it's not Kuroo's fault when it happens but it's still able to easily freeze and I don't know why yet.

This week, I'll write that famous live ebuild, keep fixing bugs and keep looking where that freeze comes from.

GSoC - Week 4 : an idea for a portage API

2010-06-21T06:47:00.000-07:00

Still in exam, week 4 is yet another half week without a lot of great things.

The biggest porting part has been done : switching to Qt 4's model/view framework.
This change will make it possible to draw progress bars in the Queue view and more easily draw icons in the main package list.

Of course, this brings some new bugs which have to be fixed but not big things.

The interesting part of this week is more about the first discussion about the portage API.
The goal is to make a Python API for portage so that portage GUIs don't have to execute the emerge/layman/equery commands and parse the - sometimes changing - command output.
From this API, I will create a C interface for other language based GUIs. This will most likely consist of a library that will run a helper python program so that the output can be parsed. All that will be transparent to the application using the library.

Why not rewrite portage in C ?
That would be a lot of work to rewrite portage from scratch and portage slowness is not due to using python, a C implementation would as slow as the python one.

I can't wait the end of my exam so that I can be 100% in that work :-)

GSoC : week 1

2010-05-31T02:19:00.000-07:00

If I need to remind you, I'm working on reviving Kuroo - a KDE-based portage GUI - this year.
The goal of the project is described in my last blog post.

Even though lots of things have been going around here (like the wonderful exams I'm having), I've been able to do something.
The first thing that must be done is actually revive Kuroo. In fact, it's currently not able to emerge a single package.
This week, I fixed some bugs which were in my way to achieve that goal and I arrived at some deadlock when emerging a package.
What is working though is the package checking part. It's able to find dependencies, packages size and packages version.

I will have more time this week, looking forward to it.

Google Summer of Code 2010 !

2010-04-26T14:35:00.001-07:00

Yay, I've been accepted in the Google Summer of Code program this year !

Google should maybe have a look at the way they inform students they have been accepted.
I received 2 emails, the first one saying I've been accepted for the "Revive Kuroo" project ((an adapted version of) the proposal will be posted here soon). So Yay, I've been accepted.

And the second one, saying I did not make it. That explains itself by the fact that one of my 2 proposals had to be rejected.

What's Kuroo ?
Kuroo is a Kde based fronted for portage - the Gentoo packaging system. The switch to KDE/Qt 4 killed the project leaving all Kde users without any portage GUI.

And what are you gonna do ?
The project will finish the (already started) port to Qt 4 and Kde 4 support will be dropped as we want light applications and avoid forcing people to install the whole kdelibs/kdebase (which is something that takes time on Gentoo).

The way the application uses portage (currently by executing the emerge python program) will be changed to using the portage API provided from the PackageKit project.
Compatibility with the new Portage API will also be added so that Kuroo can be used again by anyone !

And you'll do that all alone ?
I'll work alone on the project but my assigned mentor is Brian Dolbec (Porthole project) and Andrew Schenck (Kuroo project) will be there too to help in some possible troubles I might have with C++.

I also wrote a proposition to improve the KDE multihead support but that one did not make it.

I can't wait to start working on this project ! Lots of ideas are already in my head, for the GSoC and after !

So now, let's talk with my mentor and plan all this exciting Summer of Code.

Facebook chat

2010-02-11T00:29:00.000-08:00

Since yesterday, Facebook announced the immediate availability of it's XMPP compatible chat. (The first anouncement was done in May 2009)

You can now use Kopete to chat on Facebook, without the Facebook plugin.

First of all, you must go to the Facebook page to create an username if you hadn't yet.

Then, go the the configuration dialog of Kopete and create a new account. Choose XMPP/Jabber/Google Talk, not Facebook.

The configuration is simple :

Enter the jabber ID : <username>@chat.facebook.com

Enter your facebook password (or not, if you don't want Kopete to remember it)

Go to the Connection tab and uncheck "Allow plain-text password authentication"

Make sure that "Use protocol encryption" and "Override default server information" are unchecked

Click Next and enjoy !

L'ULg est muette

2010-02-10T12:23:00.000-08:00

Bon, ok, je ne parle pas souvent sur ce blog mais je compte changer ça.
Et je commence par me plaindre, ce qui prouve que je suis bien humain (oui, certains pourraient en douter)

L'université de Liège a équipé, depuis le mois de novembre, un de ses auditoires d'un système de Podcast. En effet, les cours donnés dans le local R3 de l'institut Montefiore peuvent être enregistrés et visualisés par la suite par le biais du portail MyULg. C'est ce qu'ils ont baptisé "Unicast".

Seulement, depuis la rentrée du second quadrimestre, les podcasts enregistrés sont bien disponibles mais sans le son. L'utilité de revoir un cours sans le son est quasiment nulle (quasiment car certaines personnes peuvent peut-être lire sur les lèvres, d'autres travaillent peut-être sur un projet de fin d'année consacré à la reconnaissance de discours à partir du mouvement des lèvres)

Quoi qu'il en soit, les quelques étudiants pensant ne pas devoir aller au cours en donnant totale confiance aux enregistrements ont pris un handicap et se voient en retard d'un cours.

On pourrait se demander ce qui cause cette absence de son. En effet, elle pourrait être aussi simple qu'un micro non branché, les batteries de l'émetteur volée ou encore un simple problème de configuration du système Unicast.
Le problème peut aussi être d'autre mesure. Le système d'encodage (fonctionnant sous Mac OS X) pourrait être buggé ou mal configuré ou même peut-être "piraté".
Une mise à jour d'une des applications sur la machine effectuant le traitement de la vidéo pourrait avoir cassé quelque chose qui empêcherait le traitement de la partie sonore.

Nous attendons de plus amples informations sur ce problème. En attendant, l'ULg a bien fait de n'équiper qu'un seul auditoire du système car des modifications matérielles pourraient devenir nécessaires.

Sur ce, je vais poser mes quelques questions sur le premier cours de bases de données à mes chers camarades qui devaient être plus attentifs que moi.

Older than I thought

2009-06-08T23:39:00.000-07:00

So, today, I had a look at what was on my old Fat32 partition on my USB hard drive and the first thing that seemed weird was that it was empty, the second thing was that it's last modification date was... well a really long time ago :

Isn't it ?

That's just little thingies in KDE 4 you sometimes meet, makes you smile and you will never see it again because you have no idea where it might come from.

GSoC is back on and Cazou -v

2009-03-25T06:25:00.001-07:00

Hey, it's been some time since I have written something here. Waw, my last blog entry was on September 17. In fact, I hate blogging, not because I don't have anything to say but because I'm rarely sure about what I'm talking about and I have some problems with english which I almost everytime remake my way.

Anyway, This also is my first entry on planet KDE, yay \o/. I added myself some time ago but never posted anything. So, to start with the beginning, I'm happy to be here :-).
I participated last year in the Google Summer of Code and was supposed to add support for VoIP for the jabber protocol in Kopete. I started by modifying libiris in kopete to support an early version of the Jingle protocol which has changed a lot since then. Now, I'm working with the Psi/Delta team to add Jingle support directly in Iris. Later, Kopete will be updated to a Jingle supported version of Iris and/or will have Telepathy support which will replace most of the protocol plugins inside Kopete. That explains why I'm not an active SVN committer since 2 or 3 months.

But enough about the past, GSoC is back and this year, I will - and I promise - be more verbose, I will tell you everything I'm doing on this summer of code (If I participate) and I will even try to explain what I'm going to do in advance.
My participation in GSoC this year does not mean I will totally stop working on Jingle, I'll just slow down things.
There are several project by which I'm interested, in different organizations. About KDE, I like the idea of presentation recording in Okular (+KPresenter ?) which would record timings and sound from the microphone (and animations in KPresenter ?).
X.Org's multi-touch implementation. Even though I never touched at X.Org's source code, that would be great to have something moving there. The XInput mechanism doesn't seem that complicated and the motivation comes from freinds being silly with their Mac Book :D.
The other ideas I suggested in other organizations were not suitable or I simply did not get any answers (a DBus service to replace the GnuPG smart card reader daemon, some ideas on openMoko which are apparently already implemented).

Like a charm !

2008-09-17T03:25:00.000-07:00

So, I got a new laptop, a Dell studio one and it works like a charm !

I also switched to KDE 4.1.1 which has some troubles with the system tray and the battery monitor but except that, it also works like a charm !

I successfully finished my exams and I'm now in my second Bachelor year, obviously, it worked like a charm too !

I have to work on Kopete now to finish up speex audio compression for Jingle and add some configuration dialog before merging it with trunk for - maybe - KDE 4.2 and hope it will work like a charm :-)

So, I think everything is working well for me currently !

Badabooom !

2008-07-14T10:33:00.000-07:00

Hey, it's been a long time since I've been writing here !
But I do have reasons :-)

First of all, I went in the south of France (near Cap d'Agde) and I thought I would have an internet connection there but it was not the case...

Second of all, I broke my laptop screen ! I was gently working on Jingle and out of nowhere, it kind of exploded. You can see it by yourself :

That's the reason I'm looking for a new laptop (I think it will be either a Dell Inspiron or a Dell Studio)

Third of all, my lover has intestines problems. He is currently at the hospital so I go visit him everyday.

And last but not least, I keep the precious time left to work on Kopete.
I got late on my timeline but I will make up for that lost time.
I still can't provide any screenshots now because, my work is only about protocol currently but you can look at that : It's one of the first thing visible from the outside of Kopete :

I will try to post as much as I can for now on.

6 left !

2008-06-05T14:08:00.000-07:00

Only 6 exams left and I become as free as the project on which I will really begin to work !!!
The last one will be on the 19th of June and I take 2 weeks of in the south of France from the 20th of June.
Don't worry, I will work there of course and get my coding job really started.
I talked with Justin, Remko and Olivier so know, I know that I will implement the Jingle state machine in Iris (the XMPP library which is used by Psi and Kopete).
Remko has already done a big job as he wrote the whole negotiation part of Jingle !
I stil have some talking to do to know exactly what is done and what is not. The code is not released yet so even me have to wait :-) .
As my GSoC is still for KDE, here's how you can get it at its latest svn revision :

svn co svn://anonsvn.kde.org/home/kde/branches/work/soc-kopete-jingle

Don't hope it will work now or even in 2 or 3 days, I'll blog here about the evolution of the branch.
Currently, it compiles, there is an old jingle directory which is from the last year's student who had the same project but did not succeed. This directory is not compiled anymore so this is just Kopete as you can find it in the KDE SVN trunk.

The first part of the job will be done in Iris.

But now, back to the exams, only 6 left, yahooooooo !

Google Summer of Code results

2008-04-21T12:52:00.000-07:00

What an awfull thing that waiting for the results, not seeing them arrive... and then, after all that work, you can see a result page on KDE organisation page. You see some projects with some names, some reminds you someone, some you can't even read, but you don't see yours. Fortunately, the scroll bar lets you know that there is still a lot that you don't see. So you go down, down, still some projects and some names, still some memories, but still not yours. The space under the scroll bar becomes more and more tiny, and you know you arrive at the end of the page and you still haven't seen your name but already 2 of the 3 projects you applied for are taken by someone else ! And at the time the scroll bar touches the other end, you see it, the one before the last one, your project, your name, and the name of Olivier Goffart ! At that time, the awfull thing it was becomes a wonderful thing, you are happy and want to tell it every one, so you blog it, you write aboute it everywhere, you even want to go outside and yell it !
Well, for those who did not get it, I'm a 2008 Google Summer of Code student and I will add support for Jingle video and Audio in Kopete. My mentor will be Olivier Goffart (It was meant to be, we both went/go [tricky grammatical construction] to the same university).
That makes me extremely happy and I can't wait to start coding that and more important, get my T-Shirt !

Waiting...

2008-04-11T05:00:00.000-07:00

So, it's been 4 days the GSoc student application writing deadline has been reached as we had 7 extra-days. Now, mentors are having a hard time reading them all, looking for something that interrests them, something that is a complete and ideas-full project. Students like me juste have to wait for the results which will be published on the 21st of April (if no changes happen).
That gives me time to study and try to pass those tiptoeing exams.
I should begin working on my GSoC project on the 26th of May. That's the same day my exams begin... So I'll have to start working really on it just after.
I have begun an open-source driver for my wireless card which is only supported by Ndiswrapper for now. I have a really big help from my boyfriend who is dissassembling an old proprietary linux-2.4 driver because of course, Inprocomm (now Mediatek), don't answer mails asking specifications.
So, I have some occupations nowadays. I'm nearly finished with my Apollo 13 book and I'm lost in the Hitch Hicker's Guide to the Galaxy :)

Start !

2008-03-31T02:59:00.000-07:00

Now that I subscribed for the Google Summer of Code 2008, I think it's better to have a blog where I can speak about everything I'll do this summer even if I'm not selected. In that case I'll certainly start helping the KDE project by correcting some bugs I might find.
I'll keep this blog up to date, talk about FOSS but also about my life (yes, I have a real life) and my studies.
So now, I still have some time to review my GSoC applications and at April 1 00:00 UTC, the ball won't be in my camp anymore...
I applied for Step, Kate and Kopete (as you can see, I like KDE).
For step, I'd like to add a game like The Incredible Machine, for Kate, a VI mode so the user who is used to vi can easily use Kate (or also other applications using Kate Part like KWrite, KDevelop) and for Kopete, I'd like to add Jingle video for Jabber.
I also wanted to apply for the gentoo linux installer but they did not like my Idea (porting it to Qt, they said KDE and Qt would take too much place on the liveCD, to be continued...) and for Rockbox but I happend not to have a rockboxable player so I did not one to buy one if the project would not a day support my Archos 404 (if you are interrested for the Archos 404 open source firmware, see www.archopen.org)
And there is my first article, others will follow with the results of GSoC.