opam admin check: Set with-test and with-doc to false by default

[kit-ty-kate]

Queued on top of #6331

Currently opam admin check sets with-test and with-doc to true by default, which in my opinion makes little sense, especially in light of #4541 as it does not take into account user behaviour at all but some hypothetical "what if someone had OPAMWITHTEST=1 and OPAMWITHDOC=1 always set in their environment".

This draft PR flips the default for those two variables to use what the actual default everywhere else is (with-test=false and with-doc=false). It also gets rid of -i which has no purpose anymore.

Instead this PR introduces three more arguments: --with-test, --with-doc and --with-dev-setup. However in light of #4541 i'm not sure this is a good idea and i don't this anyone uses opam admin check without -i currently, which makes the introduction of those replacements seem useless. What do you think?

Related to #5805

[hannesm]

I like this.

[kit-ty-kate]

After reflection i chose to not add new arguments --with-test, --with-doc and --with-dev-setup as they would be highly misleading as they would set the variable globally instead of the intended per package, the former does not make sense and is not possible to have elsewhere in opam anyway.

The PR is now ready to review

[rjbou]

Suggested change

	* ✘ `opam admin check` now sets `with-test` and `with-doc` to `false` instead of `true` [#6335 @kit-ty-kate]
	* ✘ `opam admin check` now sets `with-test`, `with-doc`, and `with-dev-setup` to `false` instead of `true` [#6335 @kit-ty-kate]

[rjbou]

After reflection i chose to not add new arguments --with-test, --with-doc and --with-dev-setup as they would be highly misleading as they would set the variable globally instead of the intended per package, the former does not make sense and is not possible to have elsewhere in opam anyway.

Agree. Plus, as there is no way to request check for a single package there is nothing else to add.