Re: [PATCH datacenter-manager v4 00/10] subscription key pool registry

["Shannon Sterz" <s.sterz@proxmox.com>]

On Thu May 21, 2026 at 9:20 PM CEST, Thomas Lamprecht wrote:
> v4 of the Subscription Registry. Addresses all review feedback on v3
> from Wolfgang and Lukas, and drops the trailing RFC wizard per the
> consensus on v3-0012.
>
> For the v2 -> v3 changelog and the original design discussion see the
> v3 cover at:
>   https://lore.proxmox.com/pdm-devel/20260515074623.766766-1-t.lamprecht@proxmox.com
>
> Notable v3 -> v4 (see per-patch notes for details):

-->8 snip 8<--

some general feedback for this series:

generally this worked as intended on my end. the map could benefit from
some more gestures in my opinion though, such as:

* doubl tap to zoom: useful as an accessibility guide
* double-tap and drag to zoom: very useful for one handed control

both of these would need extra support in the gesture controller from
what i can tell. also since for now that map is only really used in a
desktop context, imo this is not a big problem, but would be nice
follow-ups.

note that the drag to zoom gesture did not work for me when trying to
test this by pressing SHIFT+drag in the Chromium mobile dev tools.

one thing i found somewhat irretating is that clicking a remote on the
map triggers a map info card in the top center of the browser window
(only in firefox, chrome renders this card on top of the clicked
cluster). that card is sticky and even clicking somewhere on or outside
the map won't dismiss it. the only way i found it's possible to dismiss
this, is by zooming in enough on the map to make it go away eventually.
imo that can be irritating as the card will overlay other widgets in a
view. especially when several resources cluster this card can become
fairly long.

another thing that might make sense, is mentioning where the map data is
from. most other map widgets im aware of do this with a small note in
the bottom right [1,2]. this could also come in handy if someone claims
that we assert the "correct" boundaries for a country. which could be
problematic in some cases, such as the india-pakistan-china border
region. all of these countries have different and overlapping claims to
the region [3].

hope the somewhat nitpicky review is alright. except for the map info
issue, nothing here is something id consider a blocker, most of it can
easily be cleaned up or improved in (trivial) follow-ups. so consider
this:

Tested-by: Shannon Sterz <s.sterz@proxmox.com>
Reviewed-by: Shannon Sterz <s.sterz@proxmox.com>

[1]: compare, grafana's implementation:
https://play.grafana.org/d/panel-geomap/geomap-examples
[2]: compare google map's implementation: https://www.google.com/maps/
[3]: https://en.wikipedia.org/wiki/Kashmir_conflict