Hi,
On Tue, 23 Jun 2026 at 08:54, Steve George <[email protected]> wrote:
> - What's been better?
• Closing with ’Fixes:’ and ’Merges:’
• Easier to branch CI
• Team labels
• Notify team-foo members with @guix/foo
> - What could be improved?
Piouf, many things!
The current main drawback, IMHO, with the Web interface: After a first
round of comments when reviewing, then the submitter force-push an
update that modifies the line numbering and bang the previous comments
are now useless because wrongly located and worse they pollute the
reading!
To be honest about a year of Codeberg, I feel drowned. Today, I’m very
selective by some random algorithm about which Issue and/or PR I read
and I do it only time to time. Before I was scrolling almost every day
about almost everything. Why? Because today it’s less fun for me and
it requires from me much more cognitive overload.
I’m not happy with the current clients – maybe I’ve not tried hard
enough? – and interacting with the Web interface is boring to put it
mildly when it’s not painful.
Moreover, I share many comments from Maxim.
With my email client (Emacs-Notmuch), when reading, I jump from one
buffer with Magit to another buffer with shell, copy, paste, edit, move
back and forth, open another buffer with grep, etc. have side by side
discussion, shell, Git log, etc. Well, good ol’ Emacs! Then maybe
search something else using Notmuch, local tag a message, Org-Capture,
etc. Well, good ol’ Emacs! Maybe open several pages of the manual.
Last maybe queue the replies. Well, good ol’ Emacs! ;-)
And all that potentially off-line – at least without depending on the
state of my own network or on the state of some other network.
To be fair, when digging to old bugs with the Emacs interface of
Debbugs, one needs to be fully online. However, I have an helper where
I press ’I’ and it injects the ticket to my local mail box so that I can
deal with it later and offline.
Today, the current clients of Forgejo do not fit how my brain consumes
content. Therefore, it requires from me an effort to adapt. So it cuts
my motivation because I feel locked in one unique way to interact
without any opportunity to adjust for my own needs. The collateral
effect: it cuts my motivation about Guix as a whole.
For a trivial example about adjusting: Ludo told about cross-reference,
and indeed it’s nice that the timeline shows when it’s mentioned
elsewhere. However, using a decent email client, it’s straightforward
to have the shortcut press ’cw’ which extracts metadata ready to be
pasted elsewhere, for instance Ludo’s message:
Re: A year of Codeberg, thoughts?
Ludovic Courtès <[email protected]>
Thu, 25 Jun 2026 21:39:40 +0200
id:[email protected]
https://lists.gnu.org/archive/html/guix-devel/2026-06
https://yhetil.org/guix/[email protected]
Well, I’ve not invested too much into Forgejo clients and it’s clearly
possible equally, no doubt! However, when using the Web interface, one
needs to Copy Link and paste it. But then how the reader follows such
link and possibly offline? When some reader using a decent email client
just follows id:[email protected] and the local message opens.
All in all, I’m super happy if many people enjoy this kind of
collaboration with Forgejo. However, to be honest, on my side, few
things are just similar and all the rest is worse.
Somehow, I was already aware of that when discussing GCD 002 about
switching. I remember someone said: each heavy contributor reimplements
some half-backed forge on the top of their email client; maybe. Today,
I’m considering reimplementing an half-backed email client on the top of
an API. Hum? Not sure it’s better… Without saying API always vanishes.
Arf.
Cheers,
simon
PS: Codeberg as an organisation is awesome! Thanks to them for all
their effort and hard work in running such initiative.
BTW, I find interesting that Free Software projects as Guix have
chosen to adopt one workflow with enterprise-focused features
initially promoted by Github as the way to collaborate. No hard
feeling against such enterprise-focused features – being
enterprise-focused doesn’t imply it’s bad by design – only an
observation surely irrelevant.
