Sysop: | Amessyroom |
---|---|
Location: | Fayetteville, NC |
Users: | 28 |
Nodes: | 6 (0 / 6) |
Uptime: | 51:34:05 |
Calls: | 422 |
Files: | 1,025 |
Messages: | 90,554 |
Hi,
Now that we have deb822 format for specifying APT repo which includes everything within one file, the user should be discouraged from using
old styles where the GPG key is stored in a "keyring" (whether it be a trusted keyring or just a folder in another place with that name).
Far too many places mention the legacy and old styles. That includes
software docs which are maintained.
One cannot possibly open issues and convince in every repo, provided
if the software or doc-cycle are open in the first place.
A big red warning (and not error / failure) will bring a much needed
kinetic force for the change IMO, on the same lines as the warning
when using the old apt-keys stuff.
Hence, I request for making deb822 the default format and deprecating
the previous .list method with a noisy warning.
A big red warning (and not error / failure) will bring a much needed
kinetic force for the change IMO, on the same lines as the warning when
using the old apt-keys stuff.
Hence, I request for making deb822 the default format and deprecating
the previous .list method with a noisy warning.
Now that we have deb822 format for specifying APT repo which includes everything within one file, the user should be discouraged from using
old styles where the GPG key is stored in a "keyring" (whether it be a trusted keyring or just a folder in another place with that name).
However, can I point out that the pre-822 format has been around for
*27 years*: it's going to take a long time for people who are familiar
with it to internalise the new format. Introducing nagging warnings
too soon will frustrate people. We will have to be a bit patient.
I read some advice recently that suggested taking the time some thing
has existed to be a "half-life" for deprecation. So, if the pre-deb822
format has existed for 27 years and we deprecated it now, we can
expect 50% of users will have noticed the deprecation by the middle of
the century. By the year 2100, a little over 10% of users will still
be unaware or unwilling to change.