4 ms·

Mamba is a drop in replacement for conda which solves that.

by curiousgal 2y ago

akoboldfrying 2y ago
"Solves" in what sense? Dependency resolution is an NP-hard problem; if the tool you're using is guaranteed to run quickly, then either (1) there are cases it can't handle but conda can, or (2) the tool's author has just proven P=NP.
- worewood 2y ago
  It probably has a heuristic for it
  - okanat 2y ago
    It is also written in C++.
- StableAlkyne 2y ago
  The OP probably means "solved" as in "faster." The original Conda solver is just badly optimized, whether through the implementation or a poor language choice. For years, user performance concerns were written off as just the price you pay for the correctness of an SAT. Mamba was a third party rewrite that was a couple orders of magnitude faster. Users started leaving Conda for Mamba. It scared the Conda developer enough that they dropped whatever invisible back-end features they were prioritizing, and within months they added a feature to install different dependency resolution libraries, targeting Mamba first.
- taeric 2y ago
  It is NP-hard, but realistically that should not be a concern for most dependency management questions users throw at it? Just how many variables are folks imagining are in a typical dependency tree? Google claims the average number of dependencies is 35. My gut is most of these do not have too many versions, all told. Is this really big enough that you'd expect it to be the bottleneck?
okanat 2y ago
There is also Pixi from the same developer group (http://prefix.dev http://prefix.dev). It still uses the Conda ecosystem but it provides a Nix shell-esque environment as well as lock files. It basically maximizes the reproducibility of a development environment without the need for containers. No need for crazy remoting plugins in your IDE, or devcontainers. Pixi is also written in Rust so it should be at a similar performance as Mamba.

4 ms·

Mamba is a drop in replacement for conda which solves that.

by curiousgal 2y ago

Mamba is a drop in replacement for conda which solves that.

akoboldfrying 2y ago
"Solves" in what sense? Dependency resolution is an NP-hard problem; if the tool you're using is guaranteed to run quickly, then either (1) there are cases it can't handle but conda can, or (2) the tool's author has just proven P=NP.
- worewood 2y ago
  It probably has a heuristic for it
  - okanat 2y ago
    It is also written in C++.
- StableAlkyne 2y ago
  The OP probably means "solved" as in "faster." The original Conda solver is just badly optimized, whether through the implementation or a poor language choice. For years, user performance concerns were written off as just the price you pay for the correctness of an SAT. Mamba was a third party rewrite that was a couple orders of magnitude faster. Users started leaving Conda for Mamba. It scared the Conda developer enough that they dropped whatever invisible back-end features they were prioritizing, and within months they added a feature to install different dependency resolution libraries, targeting Mamba first.
- taeric 2y ago
  It is NP-hard, but realistically that should not be a concern for most dependency management questions users throw at it? Just how many variables are folks imagining are in a typical dependency tree? Google claims the average number of dependencies is 35. My gut is most of these do not have too many versions, all told. Is this really big enough that you'd expect it to be the bottleneck?
okanat 2y ago
There is also Pixi from the same developer group (http://prefix.dev http://prefix.dev). It still uses the Conda ecosystem but it provides a Nix shell-esque environment as well as lock files. It basically maximizes the reproducibility of a development environment without the need for containers. No need for crazy remoting plugins in your IDE, or devcontainers. Pixi is also written in Rust so it should be at a similar performance as Mamba.