GitHub repository is up

The official GitHub repository is up and can be found at:

For the first phase, this repository will hold the standard library source code and some other related code, like unit testing code, documentation and benchmarks. The license is “Apache License Version 2.0”. To be honest, I have personally studied licenses in my off-time for two weeks and have reached the conclusion that not only do you need to be a lawyer to truly discern 100% of the real life implications of open-source licenses, but you also need to consult with other lawyers too. So what I’m saying is that while we do like the general principles behind open-source and we want to open-source the code, we are not married or feel strongly towards any of the individual license offerings, including Apache License Version 2.0, which may be transitory. Additionally, choosing the absolute best license at this point is beyond our means.

A few first commits were made to the repository, but for now only UT code has been added. But not the interesting UT code, but the boring kind. If you wanted to show the language to somebody by code, the UT code would be a good place to start, since it is a bunch of relatively short snippets of code, each showing off and testing some language feature sometimes in isolation, sometimes testing their flow together. In consequence the UT code can be interesting. But not the one we committed. With the amount of refactoring in PT7, things broke often and in non-obvious ways, especially when it comes to function overloading. So we added about 50 new tests, all for single parameter overloading, creating a very complete coverage of numerical types so that we can have some measure of security that overloading never breaks again. Probably about 10 tests would have been sufficient, but maybe 50 is safer.

We’ll continue to add tests to the “master” branch of the repository, but probably not 50 in one go. But it is best to add enough tests at once to reasonably cover one small feature or API element at a time.

A branch has been created with the code from “CrashCourse – 004 – Building an Int“. The next 2 posts in this series will evolve a few classes closer to their final form while explaining design decisions and once this is done, the branch will be merged into “master”. After this, the real standard library classes will start to be added to the repository, one by one, as they are documented. The documentation infrastructure still needs some work. We have documentation in the source files using comments and XML, but we would also like to evolve this tried and true formula to also work with the exact same XML tags, but externally to the source code, so that you have the choice of documenting code on the spot, with the trade-off of making the code harder to navigate, or having the documentation fully/partially in an outside file.

Now that a few pieces of code are in the repository and soon more will come, we are forced to release a super-alpha version of ZIDE. As mentioned before, the goal of ZIDE is to offer a minimum golden standard of features out of the box for editing and building software using Z2, so you don’t have to resolve to ad-hoc solutions and command line. So a version of ZIDE must exist as long as there is Z2 code, and now there is.

At the begging of April, a super-early alpha version of ZIDE will be made available, meant for developers. It won’t have many features and it will be buggy, but hopefully this early release will help us to make it better based on feedback. Unfortunately, we don’t have the time or resources for multi-platform releases right now, so this first version will only be available on Windows. Starting with the second or third release, we will have a release for Linux too.