GSOC 2017/Student Application mikeman7

Contact Information

Michael Viveros, personal site, michaelviveros@gmail.com, 905-807-9537

Abstract

This project's main goal is to create a .NET Core library wrapping systemd. It will be developed in C# and published on NuGet. Features of systemd that will be wrapped by the library include starting/stopping services, showing the system status and scheduling reboots. This project will take advantage of utilizing existing solutions like Tom Deseyn's .NET Core D-bus library to communicate with systemd. I am a good fit for this project because I am a quick-learner, I have 12 months of ASP.NET experience at 2 different companies and I have a strong desire to contribute to open-source software.

Benefits to Community

This project will make it possible to access Linux systemd functions from C# code. This benefits the community since users can use a higher-level language/framework (C#/.NET) to manage their system as opposed to lower-level languages (UNIX commands, shell scripts). This makes it easier for new users who are unfamiliar with Linux development to be able to learn more about Linux by being able to use a popular framework they might be familiar with. Additionally, it makes it possible to write more sophisticated applications related to managing your Linux system since you can take advantage of higher-level .NET concepts like classes, interfaces and unit/integration tests.

Deliverables

Below are some potential project milestones with specific deliverables for each milestone. Note that the Design milestone will include research into related work (systemd wrappers written in other languages, existing .NET wrappers for other system functions like dbus) and reaching out to the community to find out which systemd functions would be useful to implement in the .NET Core library.

Weeks	Due Date	Milestone	Deliverables
1 - 2	Fri. June 9th	Design	List of at least 5 systemd functions that will be implemented with justification as to why these functions were chosen. List of additional libraries this project will use including examples of how they will be used
3 - 4	Fri. June 23rd	Prototype	Prototype implementing at least 2 of the systemd functions that uses additional libraries
5	Fri. June 30th	Test Prototype	Integration/unit tests for prototype
6	Fri. July 7th	Publish Prototype	Publish working prototype as NuGet package and add documentation
7	Fri. July 14th	Prototype Feedback	Get feedback about prototype from community
8 - 9	Fri. July 28th	Library	Add other systemd functions to prototype based on list of functions from Design milestone and community feedback
10	Fri. August 4th	Test Library	Integration/unit tests for library
11 - 12	Mon. August 21st	Publish Prototype	Publish working library as NuGet package and add documentation

Additionally, I will write weekly blog posts by end-of-day Friday containing information about my progress that week and how it fits into the current milestone, problems I am facing (including what I’m doing to fix them) and next steps for the following week.

Related Work

There is a systemd wrapper for Python called “systemd” which is implemented using C extensions for Python (Cython). It uses C libraries like sd-daemon and sd-journal to send systemd daemon notifications and to access logging from the systemd journal. It does not wrap any systemtcl functions like starting/stopping/restarting services or showing the system status.

There is a .NET dbus wrapper which this project can use to access systemd D-Bus APIs. This API provides access to a Manager object which is used to manage Unit objects (services) and the system through methods like StartUnit, StopUnit and Reboot. The .NET dbus wrapper makes it possible to access these systemd D-Bus APIs through C# code.

About Me

• Education - 5 years of schooling, finishing undergrad degree in Software Engineering and Management from McMaster University in Hamilton, Ontario, Canada
• Work experience - 20 months of co-op experience at 4 different companies ranging from an investment bank to an educational software company that makes a Learning Management System (LMS).
  • 12 months of which were writing ASP.NET code for 2 different companies and included projects like adding a database-agnostic layer to a product and making a widget to reset a student’s progress in a course
• Programming experience - created a Django app where users can answer census questions (CensusTest) and created a GIMP plugin which generates collages (CollageCreator)
• Open source contributions - I haven’t contributed source code directly but I had a great discussion in the Django users group about database models, I created an issue with a UI library that wasn’t working and I wrote a blog about how to write apps with web technologies using the Cordova framework

Why Fedora?

I have a strong desire to contribute to open-source software. For the past year, I've had my mind set on doing Google Summer of Code before I graduate. I turned down a job offer and postponed my job search just so I can do Google Summer of Code this summer. Fedora seems like a great organization to contribute to for Google Summer of Code and I've been interested in Fedora and Red Hat for awhile now. Ever since I read a book called “Under The Radar” by Bob Young (one of the founders of Red Hat), I’ve really wanted to get involved in an open-source community. The book is about how open source software can be a successful business and uses Red Hat as a running example. I learned about the “The Cathedral and the Bazaar” essay, by Eric Raymond, one of the key voices behind the open source movement, which argued that the best way to develop software is to make it as widely open-source as possible. One of my favourite anecdotes was about when Netscape released their open-source browser on mozilla.org, it did not contain any cryptography since it was illegal to “export” cryptography in the US at the time. Within 2 hours, members of the community from Europe released a version of the browser which had full-strength encryption. This example of the power and collaboration of the open-source community is why I really want to get involved with open-source software and I think the Fedora community would be a great place to do that.

I definitely plan on contributing to Fedora after GSOC if I get accepted and would love to contribute more to the Fedora .NET community and the Fedora community in general.

Why This Project?

I've outlined the benefits of this project to the community, deliverables I plan on completing and related work I've researched in the sections above. One reason I would be a good fit for this project is that I am a quick-learner. I’ve had to learn 4 large code bases from different companies for my co-ops and I’ve developed applications in my spare time ranging from a Django website to a GIMP plugin. So I will be able to quickly learn about systemd, D-Bus, .NET Core and how they can all connect to make the wrapper library for this project.

Additionally, I have 12 months of work experience at 2 different companies developing ASP.NET applications so I am very familiar with .NET and C#. One project I completed was adding an additional data layer to one of the company’s products to make it database-agnostic. This involved migrating database-specific queries in SQL to server-side C# code using LINQ expressions. This project resulted in fewer support issues for the company since they had one C# code base to maintain instead of several database-specific SQL code bases to maintain.

Lastly, my large amount of work experience has taught me professionalism and how to make a contribution. I’ve learned to ask questions after first trying to find the answer myself, to be diligent in every task I do and to be efficient with my time by being conscious of how long I spend on certain tasks and to consider alternatives after spending too long on a task. At my most recent co-op, I made a lasting contribution to the company by developing a continuous integration prototype, on my own, without being asked. I saw a problem with the existing manual process of developers having to copy over code to test servers and I was able to fix it by automating this process. I will have the same mindset when working on this project to be conscious of my surroundings and to take the time to think of ways I can contribute to improve the Fedora .NET community.

In terms of other commitments, I am travelling to Mexico, tentatively from June 7th - 11th. I can work extra hours during the day and/or on weekends to make up the time. I do not have any other commitments so I will be able to give Google Summer of Code my full time and attention.