Quoting Russ Allbery (2022-04-19 19:29:09) > Jonas Smedegaard <jo...@jones.dk> writes: > > > Now, some may argue that I am describing a case for package pinning > > here, and that *might* be true but I don't know that yet - because > > the proposed change to the system does not exist yet so I cannot > > really know that yet. Possibly the implementation will be so that I > > continuously need to check if some new non-free blobs was introduced > > and block those, instead of the current situation of not needing to > > do anything actively to keeping most possible risky "stuff" away > > from my systems. > > I feel like Debian already offers multiple mechanisms to prevent > installation or updates of packages, both specific packages and > packages by suite, archive, etc. I'm dubious that we need some > additional extra-special mechanism just for firmware, as opposed to > documenting the many and varied mechanisms we already support for > pinning old packages, disabling automated upgrades, and so forth. > > We need some way to clearly label non-free firmware packages so that > you can apply whatever installation or upgrade policy locally that you > want to apply, but solution #5 provides that by keeping the non-free > firmware in a separate archive area (which apt calls "components") to > which you can apply different apt policy.
The issue I have with option 5 is that non-free blobs are then enabled by default. Yes, I can then block it, just as I can block other parts of Debian. Difference is that non-free blobs are (generally) bkack magic that I need to trust blindfolded, whereas Debian as it exists today I can (hire someone clever to) verify what it does. I agree that we should make it easier for our users to choose to trust black magic "stuff" that they need to enable their devices. I do not think that we should impose on our users to trust black magic by default, though. I think that all non-free code distributed by Debian (be that code executed on the main CPU, and code uploaded to external devices, and code served to other people's web browsers) should be easy to use but opt-in, not (some of it) opt-out. Because we cannot reasonably know what it realy does and therefore not reasonably decide if sensible to trust or not. We can only blindly assume that "newer is better". - Jonas -- * Jonas Smedegaard - idealist & Internet-arkitekt * Tlf.: +45 40843136 Website: http://dr.jones.dk/ [x] quote me freely [ ] ask before reusing [ ] keep private
signature.asc
Description: signature
