Yesterday we went to Amsterdam to meet the Blender’s guys. Since most of us has never been to Amsterdam before, Boudewijn took us for a long walk in Amsterdam’s street (or should I say canal), which started in the overcrowded area around the central station, that we left as soon as possible to walk in more quieter area:

Then at the end of the walk we arrived at the Blender institute:

Where Ton took us on a visit of the Studio, and then we assisted to their weekly update, where all the members of team show what they have been working on, their difficulties and how to solve them. After the meeting, we went to a restaurant, as an opportunity to know each other, and to learn more on how their work, and in hope that one day Krita can be useful for them.

Tomorrow, I will go back home in Göteborg.

Krita is now focused on being a painting application

We have mentioned in our blogs entry (by me in Krita meeting 2010 – day 1 and by Boudewijn Rempt’s blog) on last week-end Krita meeting, that we had now decided for a vision oriented toward painting.

Where does that leave photography ? Well clearly, it is out. And honestly, between Gimp (especially with their work on 2.8) and Digikam, there is not really much room for an other linux photography application to prosper. Since Krita was always more oriented toward drawing and painting, and photographic features were available mostly because “we can”, and there is no high-end application for drawing and painting on linux, the logical conclusion, for us, was to focus on where we can be the best, and the most useful.

Removing photographic-specific features

The logical conclusion is to remove the features that are not useful for painting. This include many of the photographic plug-ins, like tonemapping, bracketingto-hdr, lens correction, noise reduction filters. As well as a set of artistic filters, but that are mostly useful to transform a picture in something that looks like a painting.

And anyway there are better tool for that job, like the excellent Qtpfsgui, in action below on Deventer’s mill:

I started a discussion on the subject on Krita’s mailing list, which triggered a bit of a uproar. Especially from people who have used Krita for photographic editing. Live with it, use Gimp or Digikam, or install the removed plugins from the future extension website, write your own, just do not count on us for that, we are going to be focused on other features.

An extension website

Since it would sadden me to kill forever some of those plug-ins, and also while we do not want to support photographic features, or features that are of no interest for painting, we also do not want to prevent people to have or use those features, they will simply not be part of the default distribution. We are going to setup a new website where those extensions will be hosted, hopefully with “nightly” build (more like regular build) to keep them buildable, and synchronized with git/hg, where a tag would trigger a new release automatically. In essence a revival of the krita-plugins project.

In the technical area, I and Dmitry had a long talk on how to improve the filter API, to make it both easier to write effect filters, retain performance and ensure that it is less buggy with respect to selections and masks. In meantime Lukas was teaching Vera how to implement new painting operation, so that she can work on a water color brush engine.

When it came to the UI, we talked about what to do with painting op presets preview, and it was decided that it would be more useful for the user to have a scratchpad where he can make his own testing of the current settings, rather than having a computer generated preview. Boudewijn is now working on implementing exactly that. We also discussed painting presets management, it is going to be very basic for 2.2, with just a list name and a preview (either computer generated or made with the scratchpad). And later we would like to have tags, search by tags.

And between two discussions, we were working on bug fixes, polishing features, etc… All the small details to make Krita an even better application. And now is the hack week, with Boudewijn, Lukas, Sven and me.

We decided to focus on painting, sketching, comic books and texturing. As well as making an application for high-end painters. The question of how much the digital painting should mimick real-world painting, and based on our experience of watercolor in 1.6 that was so advanced that it would simulate the drying itself, we decided that real world should be an inspiration, but that there is no point to make digital painting exactly like real world, if you want real world, you can just take a real brush and paint. But it does not mean that we should take inspiration in real physics, when it makes sense, like for color mixing.

In the end, it took us more than two hours to define the following vision:


Krita is a KDE program for sketching and painting, offering an end–to–end solution for creating digital painting files from scratch by masters.

Modelled on existing real-world painting materials and workflows, Krita supports creative working by getting out of the way and with snappy response.


It does not mean you cannot use Krita for something else, or develop plug-ins that solve a problem that does not fit the vision. But it means that we are going to be focused on implementing that vision, and that the default of the application will be oriented toward that vision. And when we have to make choices, we will look at the vision and see which decision makes more sense for the vision.

After the lunch, we took a digestive walk in Deventer streets:

Then we discussed about finding GSoC ideas, that would help to implement our vision, and concluded that a new transform tool and a good UI to access ressources was the two main ideas we needed. Then everybody went back to his computer, to fix bugs, to discuss UI ideas, and just to make Krita “the best application ever” (Vera’s tm).

Boudewijn explained us about the past of the city, and also which building was going to be demolished, and what nice house has been replaced by an ugly one. We walked a long many nice streets like this one:

Below is a church that was build by the city inhabitant that were denied access to the cathedral, now it is used as a concert hall:

Then we arrived at the market place, where a krita developer took a picture of a krita developer taking a picture of an other krita developer:

Upon our return to Boudewijn’s house the electricity was not back, but Vera had arrived, so we headed to the nearby Church basement which had electricity but no Internet connection. I first tried to work on making OpenGTL’s API nicer to use, while Lukas worked on importing ABR brush from recent photoshop version with the help of Vera, Dmitry worked on layer stack recomposition, while Boudewijn tried to fix a thread bug that seems to only happen on Vera’s computer.

Now electricity is back in the house, Peter and Sven have arrived, and we are finally connected again, and are planning to go in a pizza restaurant in the evening. And now Adam has arrived too. So we are all there now.

Qt has caught an exception thrown from an event handler. Throwing exceptions from an event handler is not supported in Qt. You must reimplement QApplication::notify() and catch all exceptions there. terminate called after throwing an instance of 'std::bad_alloc' what(): std::bad_alloc

In other situations, std::bad_alloc is replaced by the name of the exception. The problem is that if you now want to know where it happens in your program, the backtrace points to where the exception is rethrown by Qt’s event loop, which is not where the error happens.

I first had this problem a few days ago when implementing multi-layers support in EXR, since the exception name was specific to OpenEXR, I just grepped the code and deduce where the error occurred. But this is not very convenient when the error is generic, like std::bad_alloc which can be thrown just anywhere. And as it turned out by Qt itself in ‘qBadAlloc()’. The solution suggested by Maelcum on IRC is simply to set a breakpoint in the function __cxa_throw, which is a function of the C++ standard library that is actually doing the job when the keyword throw is used (at least with the GNU stdlib++, no idea if it is valid with other standard library implementation). And then you get a backtrace that point to the problem.

I thought I would share this tip in case, in some day, you find yourself with an uncaught exception in a Qt application.