pex 1.0

[wickman]

This is just minor CHANGES.rst cleanup before minting 1.0. But I'd like to take this opportunity to get feedback. Is now the right time to cut 1.0? Any features we're missing? Don't like the semver contract?

[warsaw]

I've been playing a bit with master (not wickman/1.0 atm) and it looks pretty good w.r.t. use in creating Python snaps for Ubuntu core. E.g.:

$ pex -o world.pex --python-shebang=/usr/bin/python3 --disable-cache -c world world

is a much nicer command than what I had been using.

One question comes to mind though: given that my setup.py for world has only one entry point:

    entry_points={
        'console_scripts': ['world = worldlib.__main__:main'],
        },

is it possible for pex to use that automatically if no -c option is given? That doesn't seem to be the case currently. It just means one less thing that has to be specified when building some snaps. (IIRC, you were mentioning that pex could build multiple pex files if there were multiple console_scirpts defined.)

[wickman]

I don't think so. First, you can specify any number of resolvables via the command line or via -r, so at that point things become ambiguous. Second, not everyone wants to build a pex file with an entry point specified (instead they just want pex-as-virtualenv.) Maybe it makes sense for this to be the default behavior of bdist_pex (see #99).

[warsaw]

Yeah, could be rather cool as bdist_pex. Okay, fair enough!

[Yasumoto]

+1