Fix inconsistency in dependency resolution documentation

[wiggin15]

It looks like during commit d7c33065ad, the example given for "dependency resolution" in the docs was changed from using "dummy" package names (package_coffee, package_tea and package_water) to "real" package names (pytest and pytest-cov). While the new example may be a little more understandable, this caused some inconsistencies and "bugs" in the documentation, namely:

• The wording .. pip cannot install the packages requested because they are asking for conflicting versions of pytest is not entirely accurate - these are not two different packages asking for pytest, but one package asking for pytest and pytest itself.
• There is a missing word in the newer documentation: with a version or equal to
• Most importantly, later in the documentation the "dummy" package names are mentioned again (In our first example both `package_coffee` and `package_tea` ...), and since this is the only place they were mentioned before this part, the docs are now inconsistent and confusing.

I reverted the example to use the older package names. I also switched the "dummy" versions a bit, because in one place it said package_tea 4.3.0 depends on package_water==2.3.1 and then later it said `package_tea 4.1.3` which also depends on `package_water 2.6.1` which makes less sense (older version using a newer version of a dependency).

[wiggin15]

I don't think this change requires a news entry (?). Can I make it skip the check?

[pradyunsg]

Yup. Add a news/10751.trivial file to do that. :)

[wiggin15]

Done.

@pradyunsg I just noticed the only other branch that pip has other than main deals with something in the same area: branch revert-10279-reorg-dependency-resolution-information commit e30bfac . Do you remember if the reason for reverting the doc reorg had something to do with this issue?

[DiddiLeija]

I just noticed the only other branch that pip has other than main deals with something in the same area: branch revert-10279-reorg-dependency-resolution-information commit e30bfac . Do you remember if the reason for reverting the doc reorg had something to do with this issue?

That branch belongs to #10660. Quoting from @pradyunsg's comment:

The branch name suggests that this reverts #10279.

However, it is not immediately clear what this PR does, so I'm gonna go ahead and close this out eagerly.

We don't what was the branch for. You can just ignore it.

[ichard26]

It's probably worth merging in main just to trigger the docs build and ensure it's still green, but a local build did seem to work (after merging in main manually).

[ichard26]

Thanks @wiggin15 for the PR! (and apologies for taking so long to getting it merged)