# EasyBuild v5.1.0 (blog post WIP)

EasyBuild v5.1.0 was released on 26 May 2025.

This version includes a couple of changes and enhancements we want to briefly highlight here.

For a detailed overview of all changes included in EasyBuild v5.1.0, [see the release notes](https://docs.easybuild.io/release-notes/#release_notes_eb510).

## Build summary

EasyBuild v5.1.0 will print a summary of the installations that were planned for a particular session, and indicate whether each installation was a *success*, which ones *failed*, or if they were *skipped* (for example because a required dependency failed to install).

Here's an example:

$ eb matplotlib-3.9.2-gfbf-2024a.eb -r
...
== Build succeeded for 2 out of 4
== Summary:
   * [SUCCESS] Python/3.12.3-GCCcore-13.3.0
   * [SUCCESS] SciPy-bundle/2024.05-gfbf-2024a
   * [FAILED]  libpng/1.6.43-GCCcore-13.3.0
   * [SKIPPED] matplotlib/3.9.2-gfbf-2024a

With this summary you can more easily assess the overall result of the EasyBuild session, and quickly focus on the failing installations.

## CUDA sanity check

...

## Data installations

EasyBuild v5.1.0 adds support for installing datasets as modules.

Sam Moors (VUB) gave a talk about this idea at EUM'24 (see [slides](https://users.ugent.be/~kehoste/eum24/019_eum24_datasets.pdf) + [recording](https://www.youtube.com/watch?v=13q_aKDDv9k&list=PLhnGtSmEGEQild9FmlP8Qmz9Csc_gOJKF&index=20&pp=gAQBiAQB)).

Datasets get special treatment in terms of where they are installed, and how their "source" files are managed.

For more information, see [https://docs.easybuild.io/datasets](https://docs.easybuild.io/datasets).

## Downloading pull request diff via GitHub API

...

## Cleaner output for sanity check step

...

## 2025a common toolchains

...

## Notable bug fixes

- run `pip check` only once per installation
- download progress bar
- `CMAKE_LIBRARY_PATH` in module files
- `--ignore-test-failure`
- LLVM easyblock