Google Summer of Code 2010

Forum for talking about anything else Ecere/eC related.
Post Reply
joeyadams
Posts: 1
Joined: Sat Feb 06, 2010 9:46 pm

Google Summer of Code 2010

Post by joeyadams »

Google Summer of Code 2010 is here! The window for mentor applications is March 8-12 (next week). We probably should have started this discussion much sooner.

What's there to talk about? Well, people who will be eligible to participate may want to post here (you must be a student enrolled in college over the age of 18 to participate as a student). We could also go over ideas for what needs to be done.

Here is a list of project ideas I am aware of. I have been out of touch from Ecere for a few months, so some of the information below may be outdated:
  • Refactor/rewrite the compiler. I think I'm speaking truthfully when I say that nobody (not even Jerome) really understands the compiler codebase as it stands. There's a lot of code repetition, and adding new features often involves adding code to all of the passes (a lot of that code is missing, too). Although refactoring the compiler is important for the future growth of Ecere, short-term deliverables should include a fix for the really slow compiles and a more rigorous specification of the eC language itself. As an added bonus, it would be nice to have cool new features like closures (nested functions that can access outside variables) and elegant plurals (ask sacrebleu about those). Advice: set rigid goals, and don't make promises you can't keep.
  • Make Ecere work (natively) on 64-bit platforms. This primarily involves fixing code that acts like uint (4 bytes) can hold pointers (8 bytes on 64-bit), though you'll probably have to dig deep to uproot this problem.
  • Make Ecere look nice. Ecere's current default skin looks like it's from 10 years ago (eligible students don't need to be told that ;-) ). Depending on the approach, Ecere's theme support code may need work. It would be nice to port themes like Clearlooks, Oxygen, etc. to Ecere, or draw up a theme manually. In my humble opinion, a good default theme will draw inspiration from the Ecere logo (grays and blues) while holding a lightweight yet modern appearance. Since I don't think you can get away with not coding much in GSoC, perhaps the goal should instead be "streamline the theme creation process". Such a goal might include implementing a GUI for theme editing.
I am eligible to participate as a student in GSoC 2010. I might even be able to participate as a mentor, as I've done substantial twiddling in the Ecere codebase.

Anyone else here eligible to participate in GSoC (as a student or mentor) or have project ideas?
jerome
Site Admin
Posts: 608
Joined: Sat Jan 16, 2010 11:16 pm

Re: Google Summer of Code 2010

Post by jerome »

Thanks Joey for bringing this up again this year. Man does time fly, seems like yesterday we were applying for GSoC 2009.

I will admit I am a little reticent in applying again this year. Do we really stand a chance or are the same organizations going to be picked up again? Last year I pretty much set up this page as the ideas page.

The 3 ideas you pointed out are in there, among others. Let me comment on those.
Refactor/rewrite the compiler
In fact, I do understand the compiler code-base, although at times, like any complex code, it might take me a few minutes to retrieve the proper synaptic connections to understand a particular piece of code. However, it does need some refactoring, although I don't see any value in having to start from scratch. I have some particular ideas about how to proceed with the refactoring, which I have discussed with Redj among others. They basically revolve around a more Test Driven Development (TDD) approach. The goal would be to first build up an exhaustive array of eC test scenarios. This would first of all help define more clearly those eC features which are not yet 100% decided on how they should behave in particular cases, and also identify those 'holes' in the compiler. These tests would be the basis for writing up the rigorous specification of the eC language, which they would implement. We should probably focus mainly on those eC features which are specific to eC, as the current implementation relies on existing C compilers. Once this task is accomplished, it would be a lot easier to re-factor and improve the compiler and have confidence thanks to solid regression tests. Ryan has initiated the development of a 'Test Suite' for this purpose in his "Ecere" Summer of Code internship 2009. It still needs quite a bit of work to be used in this context. The actual re-factoring would first benefit greatly from a more object oriented design of the compiler constructs, which is currently a quick port of the original C version. A lot has been learned in the process of writing up the initial version and improving it over the years, and that can definitely help laying out a good design for an improved architecture. As for the declaration reordering/topological sort issue (the #1 suspected culprit for the slow compiles), that's something I've been wanting to work on for a while but still haven't found time for it. That's indeed high on top of the list. Closures and elegant plurals? Not quite sold on those. I think we've got a whole bunch of more important things to do. Just one example: a string data type.
Make Ecere work (natively) on 64-bit platforms.
We want that.
Make Ecere look nice
Last summer Ryan started working on a new layout/GUI/styling engine system. See this topic. This is our road to making Ecere look nicer.

An other idea of particular interest is the dynamic eC compilation with libTCC, which would be a great addition to the SDK.

See the Roadmap and Development section of the forums for more ideas.

I have little time to work on the actual GSoC application process, so if you can give me a hand Joey that will be much appreciated. I might have some time over the weekend if we can be synchronized (and are able to do anything useful before the application window starts).

Cheers,

Jerome
jerome
Site Admin
Posts: 608
Joined: Sat Jan 16, 2010 11:16 pm

Re: Google Summer of Code 2010

Post by jerome »

I submitted our application to GSoC 2010!

Let's all cross fingers now.
jerome
Site Admin
Posts: 608
Joined: Sat Jan 16, 2010 11:16 pm

Re: Google Summer of Code 2010

Post by jerome »

Well folks, sorry to bring the bad news, but it seems that unfortunately we were not selected for GSoC 2010.

I'm hoping we'll still have a great Ecere Summer of Code with nice contributions :D

Cheers,

Jerome
Post Reply