Allow more truthy values for `CUCUMBER_PUBLISH_ENABLED`

I tried CUCUMBER_PUBLISH_ENABLED=1 and nothing was published. Any value should be considered truthy except /false|no|0/. This is how it’s implemented in Cucumber-Ruby.

[aslakhellesoy]

I already did. It’s at the first line in this issue. I’ll elaborate for clarity:

Environment variables were introduced in UNIX in the late seventies, and the way C code and shell scripts behave is that any value of an environment variable resolves to a truthy value. This is because the stdlib getenv function either returns a string (truthy) or NULL (falsy).

This works fine in most cases. If you want falsy, just don’t specify the variable. If you want truthy, specify any value. This means 0, 1, true, false, whatever are all truthy.

This is why I was surprised when 1 was not interpreted as truthy. It deviates from standard UNIX behaviour, and the underlying programming language is irrelevant.

The fact that all values are truthy might not be intuitive to some, but it’s the way environment variables are designed to work. Some tools have adopted conventions that deviate from the original behaviour by interpreting a small set of values as falsey. The most common ones seem to be 0, false and no.

Git for example, interprets 0 and false as falsey. 1 and true as truthy. Any other value is an error. (Try GIT_TRACE=1 git status).

For users who are using the values true and false, this change will not alter the behaviour.

Users who have specified 1 or banana will now get a truthy value instead of a falsey one.

Are you suggesting there are users out there who might have been expecting falsey for 1 and other non-true strings that will now be confused about the new behaviour?

(BTW, I agree we should not attempt to internationalise this).

[mpkorstanje]

From @vincentcapicotto: https://theburningmonk.com/2020/04/hit-the-6mb-lambda-payload-limit-heres-what-you-can-do/

@aslakhellesoy wrong issue?