Packaging and Python 2

I tried this query:

  COUNT(*) AS count
  file.project = "awscli"
  count DESC

The result is here:

Oh, “null” is the most major distro! I don’t know what environment are included in “null”.
Ubuntu 16.04 is the second most major distro. But I believe Ubuntu 18.04 will overtake it at some point.

I found very interesting comment.

For CLI v2 we’re planning on distributing standalone installers for all platforms to avoid these and other issues.

Not that I don’t share this goal, but IIRC Donald has 20% of his time allocated to OSS, so Amazon is one of the few companies paying people to support packaging. :wink: Of course, the burden spreads further than Donald, but I like to take every opportunity I can to recognize companies that are actually doing The Right Thing™ and supporting OSS. :smile:

1 Like

(Especially since py2 doesn’t ship with pip.)

( ?? ->

Oh yeah, for sure. I think it’s more than 20% even. I’m not saying we should get out the pitchforks :-). It’s much appreciated. But, just to overstate the obvious, it doesn’t mean they can stay on py2 forever while volunteers like Paul and Pradyun support them. So they still need some kind of plan.

Huh, whoops. Today I learned!

Huh, whoops. Today I learned!

You’re forgiven :slight_smile: I forgot that there is at least one good reason why ensurepip isn’t as well known in Python 2.7. Besides the fact that is was added late in 2.7’s life, we also elected to make its installation opt-in so, unlike with Python 3, you need to explicitly add --with-ensurepip to ./configure.


Is it included by default on the installer?

Edit: Yes. says:

On Windows and Mac OS X, the CPython installers now default to installing pip along with CPython itself (users may opt out of installing it during the installation process).

1 Like

I’m pretty sure it’s possible to craft a query that shows what proportion of Py2 downloads come from what packages.

If we see a substantial amount of downloads for a certain package, that’ll be an indicator here if there’s other places to look. (It’s basically the same analysis as @methane and @njs applied to aws-cli but more generally applied)