Skip to content

PIP uses the cache by default if available #5

@peardox

Description

@peardox

This can be undesirable, particularly when testing

I've managed to bodge in some hacky progress indication for my project so this is the first time I've spotted this anomoly.

I just noticed Torch was taking 50 seconds to install on my dev Laptop while a more reasonable figure is about 6+ mins (376s on testing)

There are two stages to a PIP install, aquisition and delpoyment. If the package is already in the cache then there's very little to do in the aquisition stage as the package's whl file is already in the cache.

While it's actually beneficial it's worth pointing out that if you have the same packages installed on System Python as you're trying to install for an Embedded Python then install times will be misleading when testing unless you 'pip cache purge' before running an application.

Hmm - wonder what happens if you put a fake package in the cache?

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions