Tuesday, November 24, 2009

A new branding?

KDE has just been re-branded from just a Desktop Environment to a Community. In Fact, I would say that It has just been publicly stated, and that we, all the members of the KDE community, already knew that this was more, much, much more than just applications.

So, to my eyes, and seeing the lack of response on the PlanetKDE (which either tells us that they agree or that they are shocked :), I assume that for most of the already-members' eyes; what just happened is that we left a self imposed limit behind. We are letting the world know that we are part of a community for the people. Applications count (I can't go back to Windows, everything seem so limited now), but we mostly want to be part of it for the people.

Friday, September 11, 2009

Matt

Reading PlanetKde I found out terrible news: Matt Rosewarne passed away.This is so shocking.
I met Matt at 2008's Akademy, we arrived at the same time, one day earlier, and got into the same room. We befriended very fast. Mostly because he was such an amusing and funny person.
After that, we hanged around for most of the convention together. We talked, walked, coded, sang and drank together. Heck, we even went for "tourist day" in Brussels (Charles "njaard" came with us that time).
Unfortunately, after we split up I couldn't locate him on IRC, most of the persons I talked to called him "Missing In Action".
So, today, when I read that he had passed away, it just turned my stomach. I knew he had some kind of a disease (he took quite a bunch of pills, hard to miss it), but never went into specifics. I still was expecting to find him on IRC or in the next akademy to help him buy another whip if the need arose (we had to search thoroughly to find the pink one!).

I give my most sincere condolences to his family. What else can I say that hasn't been said before?. Perhaps, that this is a loss to the FOSS community and, to me, a personal loss.

Good bye my good friend. Till we meet again.

Saturday, July 4, 2009

The completness

The completness of the Richard Stallman's conference is amazing. It has laughs, tears, drama, religion and even music. You name it.

Sunday, April 26, 2009

Why isn't there a problem to be Konquered

Recently there's been a couple of posts by Kyle Cunningham suggesting, for the n-th time, that KHTML should be dropped.
That seems to be a recurring topic in the planet, and already some answers have appeared to the most recently discussion, both of which shows that people is tired of that topic, of course by people I mean, in this case, Benoit Jacob the writer of both posts and also those who have answered the very same question in the past. Of course I must disagree with him in one central point: every opinion should be heard and as such no policies should be applied to the PlanetKde.
So in my opinion, why isn't there a problem to be konquered? well simply because Konqueror, and KHTML by extent, is being maintained. So?, that's it?, yes, basically. We can elaborate it more: Konqueror and KHTML offer the integration with KDE that Webkit lacks, and, if we can believe the opinion of a KHTML developer whom I discussed during the last akademy, actually has a cleaner codebase.
I'm not saying that the points raised by Kyle Cunningham aren't valid, but they've been answered over and over in the past. And, to be honest, this is Open Source Software after all, not because someone decide to let die a project it will actually die. In the end, those who code have the last word, and are those who decide whether to continue or drop the towel. Right now the KHTML developers have decided to continue, and actually I'm trying to join them, but I should tell you in a later post; so please stop trying to hurt the community. Either that or work in your own alternative, if such a project is proven to be more viable, then the community will stand by it. Yes, in that aspect I agree with Benoit Jacob.
So I'm left with nothing more to say other than: please stop trying to kill good, and mantained, projects.

Sunday, April 5, 2009

Another new Arch user

I wasn't going to blog about it but once I saw Rob Scheepmaker's post about his change of distribution I couldn't resist myself.
I was dragged towards the FOSS world by OpenOffice.org. It was the very first Open Source program I ran, later I installed firefox and was still a happy Windows user. Surfing through the Web I found an announcement about this thing called Ubuntu which used Linux, I always saw Linux as a geeky thing and as such I was very curious and wanted to use it. This one promised to be easy. I downloaded the Ubuntu CD and very excited I proceeded to install it, it was the 6.04 version, the 6.10 was still beta and I didn't want to take any chances.
I loved it, I mean it, I had a feeling of freedom I couldn't describe. Although a few days later I got a little bored of Gnome (I couldn't play much with it) and decided to give KDE a shoot. Again, I was so excited when I was writing those commands "apt-get install kde".
Needless to say I was blown away by KDE and reinstalled using Kubuntu 6.10 (if I recall correctly) and I had been a happily kubuntu user eversince... till a few months ago.
It all started when I upgraded to the beta of Kubuntu 9.04... everything began to feel slow, very slow, I was puzzled because I hadn't installed anything new, but I said "it's a beta, it's normal", so I disabled composition and seemed to help performance,although I loved KWin's new effects, I lived with that... then I started to have problems with the painting of windows, a lot of garbage was being painted on my windows, I said "What? this is an Intel system, I've never had any troubles with it" then I looked and yes, was a known problem for Intel-based video cards. And I thought "it's beta, it's normal"... and I lived with that.
Later when I fired up OpenOffice I had no spell checking!. Had to look and ask at IRC, yes the maintainer had forgotten to provide the right name for the package, a symlink and problem solved. And I thought "it's beta, it's normal"... and I lived with that.
Then What really pushed me beyond the edge was that I plugged a mouse device and it froze the system!, I was shocked. When it happened again I thought "it's time for a change".
So I started looking. I had the idea of using Arch for some time now and with the pain of first configuring it gone thanks to the Chakra project I went for it.
So far I've had some minor issues which I've been able to tackle. And besides that I have to customize it to use the packages I use I'm a very happy new user. It's fast and it's nearly vanilla KDE, so what else can you ask for? =D.

All I want to say is that I have nothing against Kubuntu and its fillosophy, it was there for me to make the switch and I'm sure it'll help others to do it. But please, can QA be taken more seriously?.

Wednesday, March 18, 2009

How to assert and still have your compiler happy.

I've been questioning whether to post this or not, since may be there are a more than a dozen ways to solve this problem and I might just be overcomplicating a simple meter. In the end I decided to ask and expose my ignorance because that's the only way you can learn =).

Recently, relatively of course, a new problem has arose in the KDevelop mailing list. In essence the problem lies in code like this:

bool check = foo.doSomething();
Q_ASSERT( check )

or

bool check = foo.property()
foo.doSomething()
Q_ASSERT( check == foo.property() )
You get the idea.

Which is better than what it used to be:

Q_ASSERT( foo.doSomething() )

very effectively making the project not work on release mode =D.

However the first statement, while correct, creates noise in the form of compiler warnings about unused variables. A solution that somebody proposed was just:

#ifndef QT_NO_DEBUG
bool check =
#endif
foo.doSomething();
#ifndef QT_NO_DEBUG
Q_ASSERT( check )
#endif

or
#ifndef QT_NO_DEBUG
bool check = foo.property()
#endif
foo.doSomething()
#ifndef QT_NO_DEBUG
Q_ASSERT( check == foo.property() )
#endif

Legibility suffered so it was reverted, at least it was asked to be reverted, not sure if the reversion was made.
As a solution I proposed just wrapping the variables in a Q_UNUSED macro but I got in response that that's the wrong message in the second case since they are used, only not in release mode. So the question is, how to solve the problem?.
While I want to hear what you can come up with here's a thought:
For the first case we could create a new assert, name it K_ASSERT for originality sake, and have it like this:
K_ASSERT( bool check, foo.doSomething();, Q_ASSERT(check) )

which might be expanded as:
bool check = foo.doSomething();
Q_ASSERT(check)

on debug mode and just:
foo.doSomething();

on release mode.
And for the second case we would need a macro which allows us to run code only in debug mode:

K_DEBUG_MODE( bool check = foo.property(); )
foo.doSomething();
Q_ASSERT( check == foo.property() )

which, if I'm correct, would be expanded as:
bool check = foo.property();
foo.doSomething();
Q_ASSERT( check == foo.property() )

on debug mode and
foo.doSomething();

On release mode.

All in all I need that in essence what's needed is a macro which let you separate release and debug contexts allowing you to execute code in debug but not in release mode.

After thinking it further I believe that no K_ASSERT is needed but K_DEBUG_MODE can be used instead and then the first case might be written as:

K_DEBUG_MODE(bool check = foo.doSomething(); Q_ASSERT(check) )

As a disclaimer I haven't played enough with Macros to tell whether it'll work or not but what you guys think? am I taking the long route?, is there something already thought for these cases?.