On 5/12/25 09:46, Jim Klimov via Nut-upsuser wrote:
Comments from a CET's point of view.
Hi Sam,
Not really sure what you mean here. NUT documentation is written in
asciidoc, so that it is easy to combine from several source files and
render into man pages, HTML, PDF, etc. (which does go via docbook XML as a
technical detail of asciidoc, and does result in some *roff files as a
technical detail of man page rendering, but in NUT sources/recipes we do
not directly care about either of those aspects). Allegedly there are a few
quirks with Asciidoc as well (notably there are several renderers out
there, but any semblance of a formal standard and common testing suite was
being discussed as brewing up on FOSDEM 2025), but it is pretty convenient
and light-weight once you get a hold of it.
But woefully incomplete, primarily because the makers claim one thing in
their docs, but actually deliver something else. I don't consider that
as a nut fault. However, since nut is the only one providing an
industry wide solution in the face of makers who have NDI what the
non-windblows world needs, I do fault the nut folks for not holding the
makers feet a little closer to the fire, stopping the lies and outright
BS this division of our industry is rife with.
NUT "dist" tarballs, including release snapshots, do include a copy of
generated man pages (probably in a *roff format) for the benefit of
end-users who only have a compiler and do not want to burden their systems
and build times with asciidoc/docbook/etc. tooling. So they can just build
NUT programs, `make install`, and have them nicely documented out of the
box. Those generated files are not tracked in Git.
But should be.
Full-scale builds such as for packaging are encouraged to have the full
stack in the build agent (or build root) and re-generate these documents.
This might, depending on local settings, add distro watermarks ("NUT pages
as part of OS XXX docs"), apply distro-wide build timestamp, use the *roff
version that OS is comfortable with, or whatever.
Also note that since NUT v2.8.3 we added support for `configure` options
to assign man section codes (numbers or not) for systems that do not follow
suit of Linux and BSD numbering (e.g. in Solaris/illumos, the system
commands are historically not "8" but "1m"). Previously this required
strange patch files on packager side, a burden to be revised/updated for
each NUT release; now it requires just a few configure options that can be
left in the recipe once and forever.
Another point, you are about o make a 2.8.3, but the latest debian is
2.8.0. More feet to hold up to the fire. We can't use either to their
full capability because the docs don't match what we read here. The
docs don't even seem to apply to the version they purport to be, if they
exist at all. Hence I'm pleading for docs that match what the repo
installs.
I have an APC 1500wa, now several years old. Its front panel display
has been asking for a new set of batteries for at least 5 years, but due
to my now passed wife having COPD, a 20kw kohler in the back yard has an
under 10 second startup time. So this machine runs normally for that
time period. And nut is not running, fails to start since bookworm. And
I have NDI why. The APC OTOH is doing what I bought it for.
Hope this clarifies a few points?
Thanks for reading this far, Jim.
Jim Klimov
Cheers, Gene Heskett, CET.
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis
_______________________________________________
Nut-upsuser mailing list
[email protected]
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/nut-upsuser
