Re: Timing on `Rules-Requires-Root: no` being the new default

Paul Gevers:
> Hi Niels,
> 

Hi Paul

Thanks for the feedback.

> On 29-12-2024 11:12, Niels Thykier wrote:
>>   * Do you think it is feasible for us to do this change in Trixie and,
>>     if so, do you have any preferences for timing or bug counts that we
>>     should keep in mind?
> 
> If you are committing yourself to solve the issues (as you say, one line 
> NMUs possible) within a month or two, I suggest you go ahead now and 
> don't defer. If you don't commit (also fine, you're a volunteer), I 
> rather have this wait until after the release of trixie.
> 

I find this very doable and the numbers have gotten considerable better 
since I wrote to you. We are down to 34 known bugs without a patch.

Guillem and I discussing the exact timing for the dpkg upload with this 
change. Hopefully later today, but otherwise I expect in a few days from 
now (Guillem is fixing a regression in dpkg unrelated to this). Once 
that hits unstable, I will bump the bugs to RC.

>> # Key packages
> 
> Do any of these packages have risky proper fixes? I prefer to not take 
> too much risk here, so I'd prefer the single line "fix" than something 
> that takes two or three uploads to get right. I assume that building 
> with and without the patch and comparing with diffoscope should be able 
> to proof you got a patch right, so I expect you can bring the risk to 
> near zero.
> 

For patches/NMUs that I have been making in general (key package or 
not), I have kept things simple and preferred the one-liner for anything 
non-trivial.

Also, for remaining key packages without patches, the maintainer has 
shown a strong preference for the one-liner fix for Trixie. Accordingly, 
if I end up having to NMU any of those, I will certainly use the 
one-liner fix.

Additionally, the two most likely kind of regressions would either be a 
FTBFS or make 
https://udd.debian.org/lintian-tag/wrong-file-owner-uid-or-gid?affected=yes 
light up like a Christmas tree. Neither of which will be particularly 
silent.

>>   * 157 bugs would become RC if we hit the button today.
> 
>  From a social point of view, I would appreciate it if you actively help 
> prevent autoremoval due to this change if you go forward with this 
> before trixie.
> 
> Paul
> 

Certainly.

Thanks for your answer. :)

Best regards,
Niels

-----BEGIN PGP SIGNATURE-----

iQEzBAEBCgAdFiEE9ecZmu9eXGflVYc/dA1oiINl0okFAmd2hd0ACgkQdA1oiINl
0olpBAf/RXBsl+x2wf6O8K4v44Em0DQ/O07QV8yADky6smptumcEIc3l5vhNMrnh
yqSNS1gEFzgFl0OXAp+tBQqFlcVjb3Csfv0TfCC4UOdoN83pA/7t/Ia5uk2yfHjA
3/UsRzTWWjscCfrgmeEKcXhHPIsItrCOLhTMToPO0Q6IdnNJSG2MGxEGx6ZS7yxJ
CE6KlLDcINZ1dTuqRWlgkmKceSbJsmbFvljv86P6bri7JG9PdS0U1QxmcSf/2/Yx
kPQqUlByWKakhqHH1/SwFPTwxF7OS249J+jDMYhhbcbG7kHpf49NY4k7qZU7JXi1
tc0fiKA3imBcZ+jaEQ0Lr0NMyW2i7A==
=76qN
-----END PGP SIGNATURE-----