On 2026/06/12 06:00, yaydn@protonmail.com wrote:
> > > Here is what I got from a:
> > >
> > > cd /usr/ports/www/librewolf/ && make makesum update-plist update-patches
> > 
> > don't run makesum if you're testing something that has been
> > sent out. you want things to fail if the downloaded file does
> > not match what the submitter has.
> > 
> 
> That makes sense. In this case I did a patch in parallel with Leah's. I
> got a similar diff to Leah's. Is that alright? I apologize for not
> explaining properly.

ah yes. in that case, if you're replying to an email with a diff, and
you've tested something which isn't that diff, then that's important
information to mention :-)

> I also forgot to mention building this on current/amd64.
> 
> > update-plist/update-patches can be useful if you're checking
> > whether something was missed however that would probably be
> > done after a clean build.. (and with some ports running
> > update-patches at a stage later than "make patch" can cause
> > spurious changes)
> > 
> 
> To make sure I understand, when testing someone else's patch one should
> do 'make build' or 'make package'(that includes the prior?). Then use
> update-plist/update-patches for testing for any missing changes. That
> thankfully is what I've been doing.

you could do "make patch" and then immediately "make update-patches"
to make sure that they haven't missed running it.

running update-patches at a stage later than that is not reliable.
in most ports it doesn't matter but, in some, changes are made to
patched files with sed replacements or similar. in those cases,
running update-patches later will change placeholders in patches
for the replacement text.

> I assume make build/package would implicitly do the prior steps in:
> https://www.openbsd.org/faq/ports/testing.html
> or
> file:///usr/www/faq/ports/testing.html
> if one has a checkout of the /usr/www tree.
> 
> Is that wrong; should I be explicitly doing each step or am I failing
> to do a needed step?
> 
> Separately, where as when in the pipeline to make one's own patch, is
> the workflow 'make makesum update-plist update-patches package' fine or
> should I be explicitly doing 'make build' prior to the
> update-plist/update-patches?

update-plist, yes.

update-patches, no as above.

> I thought make 'update-plist' for example implicitly did 'make build' if
> it has not been done so already. May have misunderstood the
> https://www.openbsd.org/faq/ports/guide.html
> or
> file:///usr/www/faq/ports/guide.html
> again(due to my own poor reading comprehension, it is a good guide).
> 
> Expanding a bit, if no new patches are needed in steps 13 and 14 of
> guide.html is the update-patches in an okay position for finding any
> needed changes so old patches in patches/* apply cleanly?

that's talking about writing ports where you might need a cycle of
attempt-build / patch / attempt-build / patch / etc. in that case
yes you're going to need update-patches at a later stage than right
after "make patch", and then you'll have to deal with undoing any
changes that were made that you didn't want.

> Sorry for my confusion and if my query is not well formed. Did my
> questions make sense?

yes.

> > > cd /usr/ports/www/librewolf/ && make port-lib-depends-check check-shlib-syms
> > 
> > check-shlib-syms requires a checkout of /usr/src.
> > 
> 
> My culpa.
> 
> Thanks and have a good day/night.