Hi Maytham,

Thank you for taking the time to write this, it is very helpful and much
more actionable than the feedback I had so far.

> Why are you packaging a development snapshot rather than the latest
> v0.4.0 release?

You are right, there is no good reason for it now. Back in October the
snapshot was ahead of the only release; I have now prepared the
packaging for the v0.4.0 tag, and anything newer would go to
experimental first.

> Right now, in it's current state using the latest release, is portop
> mature enough and suitable to be included in a stable release of
> Debian, where it will sit unchanged for 2 years [...]? Are you
> willing to ensure it is looked after for the full lifetime of a stable
> release by backporting any such fixes (minimal patches)?

Yes, I am willing to look after the version in stable for its whole
lifetime, backporting security fixes and important, non intrusive bug
fixes as minimal patches. The project is young and still gains
features, but each release is tested on Linux, macOS and Windows in
CI, and for a stable release I would only ship fixes, never new
features.

> Will direct users to continue using your external APT repository
> because Debian is outdated (as many upstreams do), and if yes then
> what is the point of packaging?

No. Once portop is in Debian, the documentation will point Debian users
to the Debian package as the normal way to install it. The external
repository would remain only for people who explicitly want new
releases sooner, much like backports, and for older releases that do
not have the package.

> What is the benefit of packaging portop when other implementations
> like top, htop, btop and many others are already in Debian? In other
> words, what makes portop unique?

portop is not a process monitor like top or htop; it sits next to ss,
netstat and lsof, which give a one-off listing without context. portop
shows a live table of TCP and UDP sockets with their process, the
systemd unit or container behind them (read from cgroups), who can
reach each listener (every interface, one address or loopback only)
and what the host firewall (ufw, firewalld, nftables) does with it. It
can stop the owner of a port safely, checking the process start time
first, and it is scriptable: --wait and --free for services, --diff
against a saved baseline with an exit code for cron or CI, a Docker
Compose port audit, Prometheus metrics and a Cockpit page. I am not
aware of another tool in Debian that combines these.

> Is there a reason that the package is not being maintained on
> salsa.debian.org, under go-team/packages/?

No particular reason, and I agree it is the better place. I have
prepared the packaging for v0.4.0 in DEP-14 layout (debian/sid,
upstream/latest, pristine-tar), following the team conventions, with
the Debian Go Packaging Team as Maintainer and myself in Uploaders, so
other contributors can work on it freely. It builds in sid against the
packaged libraries, and passes lintian and autopkgtest. I will ask the
Go team for access to go-team/packages and reply here with the link
once the repository is there.

While preparing it I also found that GOPATH builds fell back to old Go
defaults (the Go 1.21 ServeMux, which broke the web dashboard's routes,
and SHA-1 in TLS); the package carries a small patch for it, and the
fix will be in the next upstream release so the patch can be dropped.

Thanks again for the feedback, it was not discouraging at all.

Best regards,
Luca


Il giorno sab 14 nov 2026 alle ore 02:44 Maytham Alsudany <[email protected]>
ha scritto:

> Control: tags -1 + moreinfo
>
> On Wed, 2026-10-14 at 23:27 +0200, Luca Padovan wrote:
> > Fair enough. Thanks for the constructive feedback anyway.
> > Best regards, Luca
> >
> > Il Mer 14 Ott 2026, 23:20 Andrey Rakhmatullin <[email protected]> ha
> scritto:
> > > On Wed, Oct 14, 2026 at 11:18:04PM +0200, Luca Padovan wrote:
> > > > I understand your concerns, but I'd appreciate some more specific
> feedback
> > > > on why you wouldn't recommend sponsoring portop.
> > >
> > > This feedback won't be actionable for you.
>
> It's because the project is young (only 3 months old), and upstream and
> downstream being the same is a red flag for many DDs. Though, having a
> quick (superficial) look at the project, it looks to be of great quality
> regardless.
>
> These are the questions that need to be answered before this package can
> be considered for sponsoring, at least in my opinion:
>
> Why are you packaging a development snapshot rather than the latest
> v0.4.0 release? If you want a specific feature, why not package v0.4.0
> first and then upload v0.4.0+git... to experimental? Consider anything
> in unstable is going to migrate to testing and then later propagate to
> stable.
>
> Right now, in it's current state using the latest release, is portop
> mature enough and suitable to be included in a stable release of Debian,
> where it will sit unchanged for 2 years with nothing but security fixes
> and non-intrusive important bug fixes (i.e. no new features)? Are you
> willing to ensure it  is looked after for the full lifetime of a stable
> release by backporting any such fixes (minimal patches)? Will direct
> users to continue using your external APT repository because Debian is
> outdated (as many upstreams do), and if yes then what is the point of
> packaging?
>
> New packages must stand up to scrutiny on why they are being introduced.
> What is the benefit of packaging portop when other implementations like
> top, htop, btop and many others are already in Debian? In other words,
> what makes portop unique?
>
> Is there a reason that the package is not being maintained on
> salsa.debian.org, under go-team/packages/? Even though you are upstream,
> it is better for the packaging to be separate so that all the other
> Debian contributors can also easily make changes, especially considering
> you want team maintenance. Many sponsors are looking for a git repo, and
> Go team packages should be created following git-buildpackage / DEP-
> 14[1] format, and follow team conventions[2]. It also makes the
> packaging easier to find and work on; I was unable to find any debian/
> directory or Debian packaging files in the master or develop branches of
> your GitHub repo.
>
> Hope this is helpful. This is not meant to discourage you, just point
> out things people like myself are thinking about.
>
> --
> Maytham
>
> [1] https://dep-team.pages.debian.net/deps/dep14/
> [2] https://go-team.pages.debian.net/packaging.html
>

Reply via email to