Virtualenv bin/activate isn't doing anything "wrong". It's just not doing it how the author believes is best. Can the whole "You're doing X wrong" meme die already please?
I like the idea though! From what I see so far the author makes a good argument for using a subshell rather than setting and resetting environment variables. At the same time though, I haven't enough experience with his suggestion yet to see the downsides... and there most definitely will be. Why? Because that's programming, man. There's no panacea, no perfect way to do things, no true "best" way to do anything.
So is this the "right" way? Probably not. Better is probably more like it and using that word instead makes you seem, you know, like not cocky asshole (not that I think that about the author at all, just generally speaking).
Can the whole "You're doing X wrong" meme die already please?
You clicked through to an article you wouldn't otherwise read. The meme persists because it works, and complaining about it just makes it more effective.
(I didn't click through because I don't really care if X is doing Y wrong.)
It does work and it makes me hate it even more! Because by the time something really is wrong I'll be immune to the meme and I'll never know what's wrong.
No, this is just bad. The whole paradigm of having build setup depend on local variables in an interactive shell (which is pervasive, virtualenv is hardly the only bad actor here) is just broken inherently. It's lazy programming. It creates all sorts of failure modes -- builds can magically stop working when people update their .bashrc files (e.g. set PATH explicitly, but OOPS the build system expects to source a script and then spawn you an interactive shell), etc...
Implicit configuration is bad, OK? Just don't do it.
Edit: in fact something like /usr/bin/virt-python my-env is much nicer than /var/lib/my-app/env/bin/python. Basically like virtualenvwrapper but for Python scripts that are not run interactively.
I regularly write scripts that involve more than one virtualenv. At the moment this is trivial by just activating and deactivating envs from one script. Starting subshells is complicated because communication into the shell barely exists. Worse than that is that the shebang only supports two arguments and is heavily length limited.
Unless you are rapidly switching envs, you could launch the parts of your scripts that operate on different envs as subscripts. Then you can pipe data in and out. Can you provide an example where this does not work?
Venv is one of the pieces of sanity in the python packaging deployment world and any improvements welcome - this goes on my List if things to think hard about
It basically figures out the pid of your shell and dumps the environment variables into a file in /tmp/swork. Then you make custom activate scripts for all of your projects. It has really helped me deal with a lot of projects. It also has convenience functions for quickly cd'ing to any project.
When you want to get back to the original configuration it simply restores the original environment vars.
Yep. I do mostly Django development and my project template just automatically sets up a virtualenv in a 've' subdirectory and sets the shebang line in manage.py to "#!ve/bin/python". From there, I just run "./manage.py whatever" and I never even have to think about it.
that's a good way...since I already use Makefiles to do little tasks in my projects, I end up just setting `PYTHON=.virtualenv/bin/python` and in my targets, using `$(PYTHON)` instead of `python`. Ditto for `pip`, `nosetests`, etc.
None of these problems are the kinds of problems I've ever run into. They're entirely a series of what if and edge case scenarios in which virtualenv is not the answer, which is fine. I don't need a tool that can do things I'm never going to do.
This is also the same technique I like to use for providing configuration to applications. Basically, write YOUR_APP_ENV=prod or whatever into /etc/environment and then add appenv to PATH which tests against the machine's environment, sets all the appropriate environment variables accordingly and ends with an `exec`. Super nice way to provide shell-friendly language-agnostic configuration.
I used to use `exec env $@` at the end of the script, so that I could `appenv > prod` for diffing, but I actually prefer this articles suggestion: `exec "${@:-$SHELL}"` since you can still `appenv env > prod`, but `appenv` provides an interactive shell by default.
Perhaps that makes sense. Actually I usually type 'workon myproject' (using virtualenvwrapper, because it provides a nice standard location and activation command for virtualenvs), so maybe this could all be transparent.
This is also how all HPC computing environments (such as SLURM, SGE) do it. They simply spawn a subshell for you to work with and do whatever you want to run.
I like the idea though! From what I see so far the author makes a good argument for using a subshell rather than setting and resetting environment variables. At the same time though, I haven't enough experience with his suggestion yet to see the downsides... and there most definitely will be. Why? Because that's programming, man. There's no panacea, no perfect way to do things, no true "best" way to do anything.
So is this the "right" way? Probably not. Better is probably more like it and using that word instead makes you seem, you know, like not cocky asshole (not that I think that about the author at all, just generally speaking).