Kim Alvefur <zash@zash.se> [Sat, 05 Feb 2022 01:09:35 +0100] rev 4888
mod_bookmarks: Also link to mod_bookmarks included with Prosody (trunk)
Kim Alvefur <zash@zash.se> [Fri, 04 Feb 2022 21:22:10 +0100] rev 4887
mod_pubsub_feeds: Allow but ignore XML comments and processing instructions
This should allow but ignore them, which should be safe.
Kim Alvefur <zash@zash.se> [Fri, 04 Feb 2022 16:25:23 +0100] rev 4886
mod_dnsupdate: Fix typo
Kim Alvefur <zash@zash.se> [Wed, 02 Feb 2022 11:22:50 +0100] rev 4885
mod_invites_page: Fix templates including stray inviter name (thanks Menel)
The util.interpolation syntax `{var?placeholder}` appears to have been
intended as an if-not, but this syntax will always return either the
variable or the placeholder-template.
Kim Alvefur <zash@zash.se> [Mon, 31 Jan 2022 11:51:50 +0100] rev 4884
misc/lnav: Attempt at log format description for lnav
The default lnav behavior often misidentifies messages with the word
'error' as errors, even in debug messages. This solves that.
Kim Alvefur <zash@zash.se> [Fri, 21 Jan 2022 17:53:31 +0100] rev 4883
mod_vcard_muc: Add a summary to improve SEO
Kim Alvefur <zash@zash.se> [Tue, 18 Jan 2022 19:09:01 +0100] rev 4882
mod_measure_process: Handle unlimited FD limits
While unlikely, better type-safe than sorry
Kim Alvefur <zash@zash.se> [Tue, 18 Jan 2022 18:55:20 +0100] rev 4881
mod_measure_process: Report the enforced limit
The soft limit is what the kernel actually enforces, while the hard
limit is is how far you can change the soft limit without privileges.
Unless the process dynamically adjusts the soft limit, knowing the hard
limit is not as useful as knowing the soft limit.
Reporting the soft limit and the number of in-use FDs allows placing
alerts on expressions like 'process_open_fds / process_max_fds >= 0.95'
Matthew Wild <mwild1@gmail.com> [Tue, 18 Jan 2022 17:01:18 +0000] rev 4880
mod_http_xep227: Add support for import via APIs rather than direct store manipulation
In particular this transitions PEP nodes and data to be imported via mod_pep's
APIs, fixing issues with importing at runtime while PEP data may already be
live in RAM.
Next obvious candidate for this approach is rosters, so clients get immediate
roster pushes and other special handling (such as emitting subscribes to reach
the desired subscription state).
Kim Alvefur <zash@zash.se> [Tue, 18 Jan 2022 09:39:15 +0100] rev 4879
mod_pubsub_summary: Fix typo in example