Subj : src/sbbs3/un_rep.cpp
To   : Git commit to main/sbbs/master
From : Rob Swindell (on Debian Linux)
Date : Mon Sep 21 2026 22:21:21

https://gitlab.synchro.net/main/sbbs/-/commit/b12622a60de7b8972b6e22f4
Modified Files:
	src/sbbs3/un_rep.cpp
Log Message:
Publish the MQTT action/post topic for user posts via REP upload (#1248)

sbbs_t::postmsg() publishes sbbs/<sysid>/action/post/<subcode> when a user
posts through the Terminal Server, but a user posting the same message with
an offline reader got no such publish. In unpack_rep() every other
consequence of a post already fired (logon_posts, user_posted_msg(), the
Posted text, the P+ log entry, signal_sub_sem() and user_event(EVENT_POST));
only the publish was missing, so any consumer counting user post activity
silently undercounted by however much of a system's traffic arrives by
offline reader.

Publish from the existing !(useron.rest & UREST_QWK_NODE) branch rather than
adding a second test, so the user-versus-QWKnet-node distinction stays in one
place. useron is the actual author here, so the payload layout is unchanged
from the interactive path, with msg.to and msg.subj standing in for touser
and title. It sits after user_posted_msg() so ptoday and posts carry
post-increment values, matching postmsg.cpp.

A QWKnet node uploading a REP is network import rather than a user post and
is deliberately still excluded; that case, along with QWK hub import, NNTP
and FidoNet, is tracked separately as a feature request in #1249.

The gap dates from e94281e6cd (fairly-47-flip, 2023-01-29), which added the
topic.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
--- SBBSecho 3.37-Linux
 * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)

.