spec file: Add Requires: %{name} to subpackage

[Olf0]

… testcases, because they do depend on Patchmanager to run.

[Olf0]

Oh, I merged PR #289 and tagged the release 3.2.2 first and then remembered that I thought of this yesterday night, so I missed the chance to also put his into PR #289. 🙁

[nephros]

Well in theory you could have it installed and "use" the "app" and the "patches" without patchmanager proper, which is why I left the dep off.

I don't think the dependency is strictly necessary or useful, but have no real objection to the change, so approving.

It should not ever depend on a version of PM though.

[Olf0]

Well in theory you could have it installed and "use" the "app" and the "patches" without patchmanager proper, which is why I left the dep off.

You mean, by applying the patch files via patch utility?
And the fact that the app can always run without Patchmanager?

I don't think the dependency is strictly necessary or useful, but have no real objection to the change, so approving.

I think (see above) I now understand the »not strictly necessary« part, but »not useful« surprises me: I would consider the statement »in theory you could have it installed and "use" [it] without patchmanager proper« as academic (as "in theory" indicates).
Practically the Patchmanager Testcases app and accompanied set of Patches exist for testing Patchmanager, right?

It should not ever depend on a version of PM though.

Ack, agreed, that currently does not make any sense: While Patchmanager and Patchmanager Testcases will carry same the version number the way both packages are generated, there (currently) is no version dependency at all. Only the »not ever« puzzles me, because if a json-Patch contains metadata (to test), which requires a certain minimal version of Patchmanager to be correctly evaluated, this may change. But the again, it would simply mean that a specific test is failing.

I really do not want to simply discard your objection, especially because the test cases are "your baby".
OTOH I do want to avoid the situations like the one we observed with Vlad: "Why does starting the app do nothing except for telling that no Patches were deployed?" Without Patchmanager installed this becomes even harder to comprehend.

P.S.: An "easy way out" would be a Recommends: weak dependency, which effectively does nothing on SailfishOS. That needs rpm ≥ 4.13 (to be recognised as a keyword) and SFOS 3.2.1 already deploys v4.14.1.
But as it is effectively a no-op, I really want to make sure that this is the right thing to do and we basically agree.

[nephros]

Well in theory you could have it installed and "use" the "app" and the "patches" without patchmanager proper, which is why I left the dep off.

You mean, by applying the patch files via patch utility? And the fact that the app can always run without Patchmanager?

Precisely.

I don't think the dependency is strictly necessary or useful, but have no real objection to the change, so approving.

I think (see above) I now understand the »not strictly necessary« part, but »not useful« surprises me: I would consider the statement »in theory you could have it installed and "use" [it] without patchmanager proper« as academic (as "in theory" indicates). Practically the Patchmanager Testcases app and accompanied set of Patches exist for testing Patchmanager, right?

Academic is a good word for my "objections", which is why I just hit the merge button.

(I had a longer reply for this typed up, but it got lost in a fatfingered reload and I'm not typing it again, sorry.)