almost-UML diagram of the locale kit
This week, I finally got the plaintext catalog add on working. Then today, Oliver reviewed my work and we had a meeting on IRC. We agreed on some changes to the internal architecture of the locale kit, and also to the classes I added. Some classes in this kit have unappropriate names, and the kit was designed with zeta compatibility in mind, whereas in today’s Be-World it seems more important to focus on gettext. I summarized our meeting in this somewhat-UML diagram. It’s not really an UML object diagram as I added informations about the workflow and the tools to use when working on a localized application, and a lot of notes and remarks. I think it shows well the status of the kit: a clearly designed API, but some missing code and a lot of TODO notes. I did not fully reported the parameters of the methods and the attributes of all the classes, look at the headers if you need some information on that. This diagram is not meant as a technical reference, but as a colourful and efficient way of cleaning up my personal notes. I think I’m better at organizing my tought in a spacial way than laying them down linearly in a text. Click on the image to enjoy it at full size :)
Well another week has passed. I cannot say that to many eventful things have happened with my project. I have been working on getting all of the logic in DriveSetup to work the way I think that it should. I finally connected my development computer up to the network today though and thought I would post a screen shot. The two partitions were created with DriveSetup! Not a whole lot to look at but I pictures are always fun :)
After a month of work, it’s time to take a break and a step back to check on our progress.
And after a month what we have is a prototype of a multi-process browser.
Haiku Native Browser
Ryan and I had a dilemma:
Where to start? In fact, there is a lot to do on this project.
So we decided to start with a multi-process browser prototype.
Mid-term evaluations for GSoC are already coming…
I’m still working on the catalogs for translating applications. I got the system working and integrated it into Haiku, so now any application can be translated. However, there is still a lot of work to do. I’m now working on a plaintext catalog add on.
Catalogs are the files that store translated strings. There is a catalog add-on called “default” that is used in applications. The problem is that the files for this add on are stored in binary form (as archived BMessages), so it’s not easy to translate them.
Another week has gone by and I am glad to say that some progress has been made. Just a few minutes ago I was able to create a brand new partition from within DriveSetup. The setup is as follows: Create a new empty Intel partition map (thanks Stippi), in that map create a new partition that spans the whole disk. It works! After that I initialized the partition with BFS and can install Haiku on it.
First up, Progress. The GSoC program is around two and a half weeks in and I thought it would be a good time for a few words. First of all a big thanks to Google and Haiku for making the transition from school to summer easy. Last summer I got an internship in a C# shop. For the first part of the summer I was not only trying to get my head around the code base, I was also learning all the in's and out's of the .Net framework. I bring this up because my first couple of weeks working with Haiku has been similar in both frustration and fulfillment. Frustration tends to come from my own lack of understanding of how the system is put together, which makes the moments when things come together seem like the best few minutes of the day. The BeAPI has been fun to learn. I have done some work with Qt, and the Haiku layout management seems to follow the style. Also the BeAPI also seems to take the "less is more" approach (compared to Qt and .Net) which has grown on me daily.
Gui programming however is not the meat of the project. For that we have to dig deeper into system. The public API for interaction with the partitioning systems aka "disk_device" API has been more challenging than the interface kit. Luckily there were a couple of examples of how to use the disk_device API in the DriveSetup application. Initialize has worked for some time now and was a good place to start. My first task has been to finish implementation of primary partition creation. Stephan Assmus had checked in the start of a creation method, and from that and the Initialize, I have been able to put together a close-to-functioning create. Deletion is the next thing I am going to finish in DriveSetup so hopefully we can have at least a basic partition editor in the near future. I will try to post a blog with the current status every week, so stay tuned for more of my rambles and updates on DriveSetup!
Bryce
Today I have started to write a catalog add-on to save catalogs in plain text for easy translation. I’ve spent some time looking at the involved C++ classes, and here is what I found.
A catalog is a collection of strings, stored as <hash,value> pairs. It is used in the locale kit to translate the text in an application to the system language at runtime. When an application starts, it asks the locale roster to find its catalog and return it back. Then, each time a strig needs to be displayed, it goes trough the catalog and is translated automatically.
As you may have read recently, I’ve had to withdraw from the code drive this summer. Luckily, another student has stepped up to take over my spot. I hope his project is a success and that I’ll be able to jump back in later in the summer.
I’ve had to reorganize my priorities because I was accepted into an REU program in which I’ll do graduate style research for the summer. I’m looking to attend graduate school so this was an opportunity I could not pass up. Thankfully Matt Madia was quite understanding and was even able to locate another student to take over from me.
I am a BTech 4th year student at Indian Institute of Technology Madras, Chennai, India.
I will be working on implementing a FUSE based filesystem for BFS so that BFS partitions can be mounted natively in Linux and other POSIX operating systems.
I enjoy building systems like these where the final outcome is really interesting to watch and useful. I will be keeping the community updated regarding the progress in this blog. Happy coding all! Enjoy your summer!!
These two weeks I’ve been quite busy with other things, so the project didn’t move as much as I wanted. However, I managed to get the catalog engine to internationalize an app for the first time. It’s not a big application, just a very simple Hello World test program. And the lack of a tool for translating catalogs means I had to edit them by hand to get the translation done.
I will be working in a Catalog AddOn writing a catalog as full text for easier editing.
