X_and_newer version selectors don't work for X <- EOLd distros

[gavanderhoorn]

Just noticed this today (but it's probably been like this for some time): it would seem the x_and_newer version selector (used when conditionally including pages fi) doesn't work if x is in the list of hidden distros.

An example is the Setting up Dynamic Reconfigure for a Node (python) tutorial. The page looks empty, while in fact the contents is just hidden. The version predicate for that page is groovy_and_newer. Changing it to indigo_and_newer makes things work again.

It would seem to make sense for the page to be rendered even if Groovy and Hydro have been EOLd, would it not?

[gavanderhoorn]

I've just changed the predicate to read indigo_and_newer on the dynamic_reconfigure tutorial, so that the page renders correctly again.

The problem described in my OP is easily reproducable by just changing it to read groovy_and_newer again.

[dirk-thomas]

The recent change from #169 might have exposed this problem. @mikaelarguedas If this happens more often now with the hidden distros being updated the wiki version macro might need to be updated.

[mikaelarguedas]

Thanks for pointing it out. With Jade being EOL in May it would make sense to update the wiki macro. An audit of the wiki for lunar release may be necessary because many other pages may be impacted by the recent change

[dirk-thomas]

I think this needs to be handled before Jade being EOLed. The recently merged PR changed moved electric, fuerte, groovy, and hydro to hidden distros and this seems to affect some pages. It needs to quantified how many pages are affected. Either those can be addressed manually or the macro needs to be updated to produce the correct result even if the distro name referred to is hidden (the second option would also prevent the problem in the future).

[gavanderhoorn]

@dirk-thomas wrote:

[..] the macro needs to be updated to produce the correct result even if the distro name referred to is hidden (the second option would also prevent the problem in the future).

Personally (but I'm just a humble user) fixing the macro to always work would seem like the only acceptable way to deal with this: groovy_and_newer implies that it is a superset of indigo_and_newer.

[dirk-thomas]

Personally (but I'm just a humble user) fixing the macro to always work would seem like the only acceptable way to deal with this: groovy_and_newer implies that it is a superset of indigo_and_newer.

👍

[mikaelarguedas]

I reverted the change for now and will iterate on fixing the macro soon. I update the wiki page to use groovy_and_newer again