Please revert changes that lead downstream projects to add extra spaces

[glensc]

downstream projects like pipenv are now forced to feed extra spaces:

-charset-normalizer==2.1.1; python_full_version >= '3.6.0'
+charset-normalizer==2.1.1 ; python_full_version >= '3.6.0'

refs:

• Extra space in requirements generation pypa/pipenv#5506
• Extra space in requirements generation pypa/pipenv#5506 (comment)
• sarugaku/requirementslib@422cdae

[ptmcg]

I will do my best, but I need more to work with. Where is the parser that parses these strings? Which version of pyparsing caused this regression? If you could supply a minimal reproducible example, I'll be happy to add it to pyparsing's test suite.

[glensc]

the full reproducer:

1. install pipenv
2. run pipenv install charset-normalizer
3. run pipenv requirements

look at the output

I'm a high-level user, I know nothing about what is the track from pipenv to pyparsing.

[glensc]

actually, that doesn't create the markers, I don't know what triggers the creation of the markers like python_full_version >= '3.6.0'