We have been working hard on removing a big pain point for conda users: Slow repodata fetching. We propose a new method for fetching repodata that is significantly faster and uses considerably less memory. I've written a blogpost detailing the issues with the current state of affairs and how our approach solves these issues: https://lnkd.in/ezwxBKcR
A big pain point for the Conda Community users is the huge "repodata.json" file. It makes it almost impossible to use Conda / Mamba or Pixi under bad network conditions. We finally found a great way to fix this: sharded repodata. With sharded repodata, the client only downloads what it needs (the tiny shards). This makes for a smaller download and a very fast initial experience, even with a cold cache. There are many other benefits (improved caching, better CDN story, ...) that will dramatically increase the quality of life for conda-forge users. Our goal is to make this available to all conda-users. We have written a very detailed Conda Enhancement Proposal and we are already working with Anaconda, Inc. employees to bring this to "conda-index" ?? Read more in Bas Zalmstra's blog post: https://lnkd.in/dif_vvvk
