Description of problem:
Missing HTML IDs (at least) in Vertical Navigation. Regression against 6.4.
Version-Release number of selected component (if applicable):
Sat 6.5 snap 1
Steps to Reproduce:
1. Login the Satellite
2. On the homepage, inspect Vertical Navigation's HTML
The tags don't have ID specified
The tags should have ID specified (preferrably the same as in Sat 6.4)
It would be good to list some ids that are missing explicitly. I suppose it will be related to https://github.com/theforeman/foreman/pull/5898
Looking at diff between 6.4 and 6.5, I see
6.4: <a id="menu_item_fact_values" data-id="aid_fact_values" href="/fact_values"><span class="list-group-item-value">Facts</span></a>
6.5: <a href="/fact_values"><span class="list-group-item-value">Facts</span></a>
I think this should be restored in 6.5 timeframe and given how many tests this break for QE, it should be addressed ASAP.
I'd be more careful with word Regression here, it's not something user would notice, so rather removing. Please add back if you disagree.
Assigning to Ohad as this BZ is related to Gilad's work on the vertical navigation.
Created redmine issue https://projects.theforeman.org/issues/26052 from this bug
Moving this bug to POST for triage into Satellite 6 since the upstream issue https://projects.theforeman.org/issues/26052 has been resolved.
FailedQA with Sat 6.6 snap 20.
Top level menu items don't have IDs. For example Infrastructure:
<li class="secondary-nav-item-pf is-hover list-group-item">
<span class="pficon pficon-network" title="Infrastructure"></span>
I don't think this BZ should be considered blocker by now.
I don't think moving this to ON_QA makes sense. Setting back to ASSIGNED because my concerns haven't been addressed.
To answer your question: While this is not an automation blocker anymore, I think this should be implemented for multiple reasons:
1) The IDs were there before and there is simply no reason to remove them - we should avoid changes that don't have any reason
2) It will make our automation cleaner and our jobs easier
3) I think major UI items should all have their IDs
4) For the sake of consistency
Ad 1) imho you can't call change of internal html ids a regression, that's not a feature or even an API. Not sure what you base your assumption on, but the reason was the reimplementation in different stack, which simply couldn't use helpers from the old one. That is a good technical reason to me.
Ad 2) that is a good justification, could we change id values? Can you list what elements you need to have ids? We'd like to keep the freedom of changing internal html structures, but we'll try to make it as little intrusive to your tests as possible. Ideally you'd let us know right after some change like that is merged.
Ad 3) we need to define some guidelines, I have no idea what "major UI items" means. Is that every div? Everything that's clickable? Everything thay contains text value?
Ad 4) could.you please better describe this? Consistenxy with what?
I'd say that anything that is clickable should have its ID. In some cases, other elements having ID would be useful, too, but it's hard to describe generally. We can talk about it sometime. Ad 4), menu items on non-top level have IDs so it doesn't make sense that top level items don't.