The first major release comes with some breaking changes to the Result component, the MultiCheckboxFacet markup, and the Facet CSS class, so be prepared if you’ve used the project before. Other than that its developers got rid off a few bugs and worked on accessibility issues.
TensorFlow users might want to update to the now available version 1.13.2 of the library for numerical computations. The bug fix release brings updates for two libraries to mitigate a number of vulnerabilities.
Go devs can try proposal
Even though it’s still some time until the next phase in the v1.14 release cycle starts, the team behind the Go programming language has marked the try proposal introduced in June as declined. The addition was supposed to improve error handling, but discussions over the last few weeks showed that many devs didn’t think it was addressing a worthy cause and that some issues weren’t considered in the proposal.
Robert Griesemer, one of the language designers, wrote in a statement that the Go team “still believe that error handling in Go is not perfect and can be meaningfully improved”, but also acknowledged that not enough attention was drawn to initial discussions of the problem. To make sure the right aspects are addressed moving forward, Griesemer asked for feedback on the most problematic ones and the impact a “good solution would have”.
LinkedIn opens Brooklin sources
LinkedIn, which is part of the Microsoft family, has taken a page out of its parent company by open sourcing Brooklin. The “distributed system intended for streaming data across multiple different data stores and messaging systems with high reliability at scale” is available via GitHub (also Microsoft) under a BSD 2-Clause “simplified” license.
According to the release announcement, the company started the project “to address our growing needs for a system that is capable of scaling both in terms of data volume and systems variance”. The project isn’t LinkedIn’s first foray into open source, though projects like Cruise-control tend to be lesser known than those of Microsoft itself.
Speaking of which, Microsoft has bumped its cross-platform machine learning framework for .NET developers ML.NET up to 1.2. The new version is meant to be able to work with TensorFlow and ONNX models and comes with TimeSeries support for forecasting and anomaly detection.
Apart from that, the update features previews of a new ML.NET CLI, and a package to integrate ML.NET models in web or serverless apps.