On 5/16/26 9:17 AM, Jakub Klinkovský wrote:
On 15.05.26 at 19:38 (UTC+0200), Robin Candau wrote:On 5/15/26 6:23 PM, kpcyrd wrote:I guess at this point of the integration I would be fine with just monitoring the developer dashboard on archlinux.org myself, without needing extra emails. That would also solve the problem of stale notifications when a co-maintainer updates the package before I even see the out-of-date email. When a user explicitly flags I'd be fine with receiving one.[...] - Monitoring the developer dashboard on archlinux.org yourself without opting-in to the notification is indeed a possibility. Alternatively (and also as a reminder for everyone else reading this mail), bumpbuddy data are publicly exposed on a web dashboard [3] and as json [4]. The former allows for a graphical & global monitoring of data with easy filtering (e.g. searching for your nick in the search bar and sort the "Up to date" column) and the later can eventually serve as an endpoint for API calls / custom scripts.How often is the out-of-date info updated on archweb? Are there some webhooks set up or does it run a job periodically? In case of the latter, I'd expect a rather long delay between updating a package and clearing the out-of-date flag on archweb, so monitoring the dashboard would not prevent the problem when "a co-maintainer updates the package before I even see the out-of-date package".
I will let Jelle correct me if needed but, as far as I'm aware, out-of-data from bumpbuddy are imported in Archweb every 2 hours by a periodic job. However, Archweb only uses bumpbuddy's data to flag the package, not to clear the flag. The flag on Archweb is still being cleared automatically / instantly when the package is released, the mechanism didn't change on that front. I'm also assuming that there is some logic to prevent stale / outdated bumpbuddy data to re-flag the package wrongly until the next run of the periodic job.
So a little delay between the package being detected as out-of-date by bumpbuddy and the package being flagged accordingly on Archweb is likely expected. But flag clearing should still be as fast as it is today.
Again, I let Jelle correct me if needed :) -- Regards, Robin Candau / Antiz
OpenPGP_0xFDC3040B92ACA748.asc
Description: OpenPGP public key
OpenPGP_signature.asc
Description: OpenPGP digital signature
