Klemens Nanni <[email protected]> writes:

> On Sat, Nov 27, 2021 at 06:29:57PM +0100, Omar Polo wrote:
>> Klemens Nanni <[email protected]> writes:
>> 
>> > On Fri, Nov 26, 2021 at 04:22:18PM +0100, Omar Polo wrote:
>> >> please find attached a tarball for pounce:
>> >> 
>> >> % pkg_info pounce
>> >> Information for inst:pounce-3.0
>> >> 
>> >> Comment:
>> >> multi-client, TLS-only IRC bouncer
>> >> 
>> >> Description:
>> >> pounce is a multi-client, TLS-only IRC bouncer.  It maintains a
>> >> persistent connection to an IRC server, acting as a proxy and buffer
>> >> for a number of clients.  When a client connects, any messages
>> >> received since it last disconnected will be relayed to it.  Unlike
>> >> some other bouncers, pounce uses a single buffer for all IRC messages,
>> >> which acts as a queue from which each client reads messages
>> >> independently.
>> >> 
>> >> Maintainer: Omar Polo <[email protected]>
>> >> 
>> >> WWW: https://git.causal.agency/pounce/about/
>> >> 
>> >> 
>> >> It's from the same authors as net/catgirl and works in a similar way.
>> >> I'm using it on a i386 box and works like a charm, it was way easier to
>> >> wrap my head around it than with znc.
>> >> 
>> >> OK/comments/feedbacks?
>> >
>> > OK kn
>> >
>> > You could build and install the extra/ tools via post-build and
>> > post-install.  Would be even nicer if upstream provided an extra target
>> > or so building these, so you'd eventually have to just set
>> > ALL_TARGET='all extra' or so in our port to ship them.
>> 
>> Agreed, I don't think I'll use them but they seems useful.  I cannot
>> even test them because they don't accept self-signed certificates and
>> I'm running the bouncer in my LAN.
>> 
>> Here's an improved tarball.  I'm calling the configure scripts for the
>> extras in post-configure, otherwise too many variables are unset and the
>> port become messier.
>> 
>> still OK? :)
>
> Hand-rolled build/install targets should probably be modeled after what
> the real targets do, i.e.

I've started with that initially, but then trimmed some variables out to
try keeping the lines short.  I do agree thought that consistency is
better than saving a line or two, and also some of the "simplifications"
I did weren't really a good idea I guess (like MAKE vs MAKE_PROGRAM.)

> [...]
>
> If upstream would provide targets for this in the global Makefile, this
> would all be obsolete and fall under a single `make' for us porters, so
> probably worth persuing.

I'll try to talk with upstream, if anything for patching the extras to
accept self-signed certificates like litterbox and catgirl do.

> OK kn either way.  Your targets work, albeit being inconsistent with how
> bsd.port.mk usually does it.  Pick what you want or leave the extras
> out, I have only tested pounce(1) itself.

since we've only tested pounce itself I'm changing my mind and I'm now
inclined to avoid installing the extras, I don't feel comfortable
committing stuff nobody tested and we can just as well add them with the
next update eventually.  I'll think about it a bit still and import them
tomorrow if nobody objects.

Thanks :)

Reply via email to