Bug#1055767: release-notes and sphinx: incompatible formatting of manpage links in bookworm and trixie
Package: release-notes
Now that the release-notes for trixie are based on and built with Sphinx /
restructedText, we got an issue with manpage links.
The manpage links are created with the Sphinx extension "extlinks" with a line
like this
extlinks = {'url-man-stable': ('https://manpages.debian.org/trixie/%s', '%s')}
in source/conf.py.
A sentence like
"Further information about this are in :url-man-stable:`rsyslog.conf(5)`. "
leads to html output
"Further information about this are in rsyslog.conf(5). "
where "rsyslog.conf(5)" is a link to https://manpages.debian.org/trixie/rsyslog.conf(5).
extlinks has just introduced a change, that makes it impossible to build the
release-notes on both bookworm and unstable systems, and getting the same
result with both builds.
Above example is the state of the r-n we actually have in git on Salsa, and
the gitlab CI pipelines running via Salsa are happily building like shown above.
But now looking at the Debian website, for example
https://www.debian.org/releases/testing/release-notes/issues.en.html#rsyslog-changes-affecting-log-analyzers-such-as-logcheck
you find the link to manpages.debian.org crippled, showing "%srsyslog.conf(5)"
as link text.
This is because *this* build of release-notes is happening on wolkenstein,
which is currently a Bullseye system.
wolkenstein will of course upgraded to Bookworm some time, but when Trixie
is released, wolkenstein will still be a Bookworm system, and the Sphinx
version in Bookworm suffers from the same issue as today under Bullseye.
So the situation is: with the current source code we cannot have the release-notes
correct on the Debian website and at the same time have succeeding pipelines
on Salsa, when people are doing changes.
(If we change the source code in a way, that the links are showing up ok
on the Debian website, every r-n pipeline on Salsa will fail. See
https://salsa.debian.org/ddp-team/release-notes/-/commit/281b371e731f8abd4a0ea755eea8165e1991bad0/pipelines?ref=master
for an example of such a failure.)
Need to investigate, how to proceed here...
Holger
--
Holger Wansing <hwansing@mailbox.org>
PGP-Fingerprint: 496A C6E8 1442 4B34 8508 3529 59F1 87CA 156E B076
Reply to: