Git Merge 2020 | March 4th | The Majestic Downtown, Los Angeles

Session

9:15am | March 4

Sno: Building on Git for Versioning Datasets

Git and distributed version control systems have made advanced development workflows the norm, allowing independent developers, teams, and organizations to collaborate on software. At the same time, we are drowning in data, and for professional data wranglers in many industries, “auckland-mar2020-final-2-fixes-1-final.zip” is still the go-to for data versioning. In this talk, Robert Coup will discuss how Koordinates built an open source version control system on top of Git that's aimed at spacial databases (0-50GB) and their users. Attendees can expect to learn how to combine Git features to work with large repositories, how to use Git as the plumbing layer inside another tool, and more.

Session

2:45pm | March 4

Stump the Experts

Session

9:45am | March 4

NoteDb: An adventure where Git is your database

Git makes a great general-purpose database because it provides significant benefits in consistency, reliability, and audit logging. But it's probably nobody's first choice for a database, and turning it into one presents unique challenges. Hear why Patrick and his team decided to use Git as their database for the Gerrit Project (an open source code-review system) and the challenges they ran into along the way.

Session

10:15am | March 4

AGit flow (Alibaba centralized workflow and command line tool)

Alibaba implemented a centralized Git workflow called AGit flow, which includes a series of patches on Git-core, a service-side filter (a new Git hook) and several server-side APIs, and a command line tool on client side. They plan to contribute the patches to Git, and open source the command line tool on GitHub in early 2020. The basic idea of the AGit flow is inspired by Gerrit, but with some differences. Hear more about AGit flow and how you can apply Alibaba's learnings to your workflow.

Session

11:00am | March 4

The Zen of Git

Git is a version control system that's widely used and often misunderstood. Is it possible to really understand how it works, and maybe even enjoy using it? In this talk, it's Tianyu's goal to get you there. This session is for anyone who has used Git before and would like to learn more about how it works, so they can have a less fearful and more productive time using it in their day-to-day workflow. Using hand-drawn illustrations, Tianyu plans to talk about the internals of Git, including how objects are stored, and how things like branches and tags fit into this representation. Using her illustrations as a foundation, she'll dive further, making commonly used commands easier to understand and implement.

Session

11:30am | March 4

Scaling the Merge Machinery

An overlooked dimension of scaling Git is dealing with large numbers of renames. Refactoring workflows can result in many renamed files—particularly when top-level directories are renamed—and this can pose a challenge for Git's on-the-fly rename detection. Palantir Technologies’s efforts to optimize rename detection uncovered multiple different strategies to tackle the problem, and also turned up a few opportunities to optimize the merge machinery for cases without any renames. In this talk, Elijah Newren will provide a high-level overview of the merge machinery and how it is used, the ideas and insights being leveraged to optimize these codepaths, and some early results from the optimization work.

Session

12:00pm | March 4

Git at Scale for Everyone

Repositories never get smaller. As your team grows, your repositories grow even faster. Git is growing, too. Luckily, the community is improving Git's performance to tackle the many dimensions of scale. Microsoft supports the largest Git repositories—Windows and Office—so Derrick Stolee’s team has a unique perspective on Git at scale. In his talk, he’ll share strategies that keep Git fast, including how to design your engineering system to take advantage of tools like sparse-checkout, fsmonitor, and partial clone. When combined with a "microservices in a monorepo" design, your team will feel like they are working with a small repository while contributing to a large project. Derrick will also share his tools to help Git manage large repositories, so your team can stop waiting on Git and can get back to what you do best.

Session

1:30pm | March 4

Virtualization for Monorepos

At ESR Labs AG, Christian Schilling and his team have implemented a novel approach to the monorepo strategy. Their tool acts as a proxy between a Git server and its clients, and is capable of transforming history on demand, enabling users to pull from and push to smaller virtual repositories. This helps to ease the transition to a monorepo from other existing setups. Christian's presentation explains the motivation behind the development of Josh and shows how his team uses the tool to solve many problems common in growing organizations and repositories.

Session

4:00pm | March 4

Storytelling with Git rebase

Session

9:00am | March 4

Welcome / History & Future of Git

Session

2:15pm | March 4

Lightning talk - Hack the Project Onboarding Process: Learning by Writing

Poor onboarding documentation can make it difficult for new developers to understand and contribute to an open source project. Inspired by her own experience bootstrapping on a Git project, Emily Shaffer’s talk will cover strategies for improving this documentation—making onboarding more efficient, instructive, and rewarding.

Session

2:45pm | March 4

Stump the Experts

Session

2:45pm | March 4

Stump the Experts