Allow adding a custom string to pip's User-Agent via an environment variable

[conspicuousClockwork]

Addresses issue #5549.

[cjerdonek]

Out of curiosity, why did you choose the word "provisional"? To me, something like "custom" or "user-data" or "user-provided" would convey the meaning more clearly.

[dstufft]

FWIW this will break PyPI's ability to parse these user agents, which might be OK.

[cjerdonek]

You mean PyPI breaks on new / unknown keys in the JSON? Is that intentional?

[dstufft]

Eh, I misread the PR because I had just woken up. I thought it was just appending data to the end of the UA, adding rando keys to the User Agent should be fine.

[conspicuousClockwork]

Out of curiosity, why did you choose the word "provisional"? To me, something like "custom" or "user-data" or "user-provided" would convey the meaning more clearly.

Back when I made the PR there were a few points made about not replacing the entire user-agent header, so the intent was to provide a way to add text in provision to the original data in the user-agent header. In any case, I don't believe there's any particular reason to keep the word "provisional". I'll update the PR with any preferred verbiage @cjerdonek.

[BrownTruck]

Hello!

I am an automated bot and I have noticed that this pull request is not currently able to be merged. If you are able to either merge the master branch into this pull request or rebase this pull request against master then it will be eligible for code review and hopefully merging!

[pradyunsg]

I like "user-data". If we're adding this, it should get documented.

Idk if it'll make it into the BigQuery dB as is but it'd be best to include that in the documentation too.

[conspicuousClockwork]

@pradyunsg PIP_USER_AGENT_USER_DATA? And as far as documentation goes, is there anywhere else that this needs to be documented besides news?

[cjerdonek]

I like user-data and PIP_USER_AGENT_USER_DATA, too.

For documentation, it doesn't seem like there's one section in the docs listing all the environment variables that are supported (aside from environment variables that correspond to command-line options). Since this is a pretty specialized variable and was related to your proxy server use case, maybe it can be added as a sentence at the end of this short section:
https://pip.pypa.io/en/stable/user_guide/#using-a-proxy-server

Alternatively, maybe a section called "General Environment Variables" could be added after "General Options" here: https://pip.pypa.io/en/stable/reference/pip/#general-options
That could list environment variables that apply to pip's operation as a whole that don't have a corresponding command-line option. (This variable doesn't seem worth also adding as a command-line option, since that adds extra maintenance overhead.)

[conspicuousClockwork]

@cjerdonek @pradyunsg is the additional documentation something that can be added to this PR or is that a separate issue?

[cjerdonek]

The same. Documentation of a change should go in the same PR as the change itself. It probably only needs a sentence or two.

[cjerdonek]

A couple comments in advance of the docs change.

[conspicuousClockwork]

Am I missing something here or is the pipeline broken? I can't find anything in the build logs indicating a direct issue with my changes.

[cjerdonek]

I opened a PR to fix it, which is why I wasn’t able to do anything here.

[cjerdonek]

@conspicuousClockwork Can you rebase your changes now that CI has been fixed?

[cjerdonek]

Thanks for your work on this, @conspicuousClockwork!

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.