If you move the project home it leaves the virtualenv folder. This is a problem if you create a new project home with the same name/path because it will use the virtualenv folder with the packages of the other project.
Describe your environment
- Ubuntu
- Python version: Python 2.7.6
- Pipenv version: pipenv, version 9.0.0
Expected result
Keep the virtualenv when you move/rename the project home. A new virtualenv when you create a new project in the same path as an older project.
That is, move virtualenv folder when you move/rename the project home.
Actual result
After moving/renaming the project home pipenv creates a new virtualenv and it has to reinstall all packages again. The old virtualenv remains there wasting space.
Creating a new project in the same path keeps the older project packages.
jcarmena
All 6 comments
then you just create a new virtualenv, not a big deal
kennethreitz42
on 18 Dec 2017
We have 100 of scripts = pipenv projects and new ones comes in and others are moved or are removed. In a year we will have 1K of venvs on all our prod machines, many of them use exactly the same dependencies. i agree it is not a big deal as it out of scope of this project. but it seems to be a big deal to create package manager for python that make most ppl happy.
s-trooper
on 23 Feb 2018
It is not a big deal to work with pyenv+pew+pip+requirements.txt either ("do one thing and do it well"). I just noticed an error prone behaviour (and perhaps dangerous).
jcarmena
on 26 Mar 2018
@s-trooper @jcarmena - I believe setting PIPENV_VENV_IN_PROJECT=1 in your environment would address this issue
jtratner
on 26 Mar 2018
emh sorry for the spam, my post was about the state of affair in python not about this fine project. in js (npm), java (maven & co.) and .net (nuget) for X projects with numpy==1.2.1 dependency i would have "numpy==1.2.1" one single time in global repository, for Y Projects with "numpy==1.6" the same. even diffent target runtimes are handled by those tools. i see virtualentiroments as a wrong way to go, its a workaround of python runtime limitation a sort of GIL for dependency management. and like for GIL, there are no solution possible at time.
s-trooper
on 26 Mar 2018
@jtratner thank you, that is enough for me. I'll set it by default.
jcarmena
on 26 Mar 2018
Related issues
johnjiang
路
3Comments
erinxocon
路
3Comments
jakul
路
3Comments
AkiraSama
路
3Comments
leileigong
路
3Comments
Most helpful comment
We have 100 of scripts = pipenv projects and new ones comes in and others are moved or are removed. In a year we will have 1K of venvs on all our prod machines, many of them use exactly the same dependencies. i agree it is not a big deal as it out of scope of this project. but it seems to be a big deal to create package manager for python that make most ppl happy.