On 2025/05/02 12:21, Peter Hessler wrote:
> On 2025 May 02 (Fri) at 11:53:19 +0200 (+0200), Kirill A. Korinsky wrote:
> :On Fri, 02 May 2025 11:28:47 +0200,
> :Stuart Henderson <stu@spacehopper.org> wrote:
> :> 
> :> On 2025/05/01 21:56, Josh Grosse wrote:
> :> > For consideration.
> :> 
> :> > diff --git usr.sbin/pkg_add/pkg_add.1 usr.sbin/pkg_add/pkg_add.1
> :> > index 67be4b80179..17b3fc12687 100644
> :> > --- usr.sbin/pkg_add/pkg_add.1
> :> > +++ usr.sbin/pkg_add/pkg_add.1
> :> > @@ -109,7 +109,9 @@ with
> :> >  .Fl u
> :> >  whenever possible, but
> :> >  .Fl U
> :> > -can be much faster.
> :> > +can be much faster, and is intended for use when running -current, if
> :> > +a new package needs to be added without conducting a full update of all
> :> > +packages.
> :> >  .It
> :> >  Replace existing packages with explicit other versions, using option
> :> >  .Fl r .
> :> 
> :> -U is a bit of a hack and I think mostly useful for ports developers.
> :> 
> :> Diff below gives a better reason *not* to use it.
> :> 
> :> Perhaps pkg_add should reject -Uu, afaict it is meaningless, and there
> :> is some meme which won't go away about running -Uu for updates.
> :> 
> :> Bonus -u -> .Fl u fix.
> :>
> :
> :Cleaner wording, indeed!
> :
> :Meanwhile, I was able to backtrace the source why so many people uses -U. I
> :don't belive that all of them read man page and decided to speedup things.
> :
> :Long story short: I almost sure that the root cause this page:
> :https://www.openbsdhandbook.com/upgrade/

Very possibly, I've run into bad advice from there before. It's almost
the new c*l*mel. Though -Uu gets repeated a lot on web forums etc, fixing
that page won't make it all go away.

> Their sysmerge recommendation is also broken, and the OS install section
> is suspect.  The networking section is also bad, most especially the
> Wireless section.
> 
> Yuck, I hope this site gets fixed.

Really there shouldn't be a need for sites like that, but that may
require some reorganisation/rewrites of parts of our docs..