On Feb 03, 2014, at 12:11 PM, Manuel de la Peña wrote:
>I have talked with barry and the check is an operation with several
>steps and adding such a feature (cancelling/stopping the check) might
>take some time to land. Is not impossible and we should do it, is just
>hard and will take time and a possible plan B should be considered.
Checking for whether an update is available is a complicated, multi-step
process. For example, there are potentially 4 gpg keyring files that must be
present, but only one that is shipped with the image, so the "check for update
available" process must be able to backtrack and download any missing keys.
It's surprisingly complicated!
The biggest challenge with adding a stop method during update available checks
is that all the various combinations and backtracking must be tested, so
that's a lot of additionally complicated test code to add. It's not
insurmountable -- I just have to find the cycles to tackle it.
On Feb 03, 2014, at 12:11 PM, Manuel de la Peña wrote:
>I have talked with barry and the check is an operation with several stopping the check) might
>steps and adding such a feature (cancelling/
>take some time to land. Is not impossible and we should do it, is just
>hard and will take time and a possible plan B should be considered.
Checking for whether an update is available is a complicated, multi-step
process. For example, there are potentially 4 gpg keyring files that must be
present, but only one that is shipped with the image, so the "check for update
available" process must be able to backtrack and download any missing keys.
It's surprisingly complicated!
The biggest challenge with adding a stop method during update available checks
is that all the various combinations and backtracking must be tested, so
that's a lot of additionally complicated test code to add. It's not
insurmountable -- I just have to find the cycles to tackle it.