On Fri, Apr 23, 2021 at 05:03:04PM +0200, Andrea Bolognani wrote:
This makes the output more compact by grouping together all
images that are built on the same base OS.
Yes, it definitely does make the output more compact, but I'm still not
completely sold on the idea that this is an actual improvement...
Later on, when we change the actions that operate on
container images to accept an lcitool-style --cross-arch
argument instead of expecting the name of the image to have
the cross-building architecture baked in, this will allow
users to quickly copy-and-paste the necessary information.
...you can quickly copy-and-paste the image name even now and I would argue
that it's even quicker than after this series:
./helper list-images
./helper <action> <copy-paste image name> vs
./helper list-images
./helper <action> <copy-paste image name> --cross-arch <copy-paste cross
arch>
Right now you wouldn't even need the verification code introduced in patch 12
since you have to paste the image verbatim which IMO for this utility helper
that only serves a very specific use case of a single repo is "good enough".
I guess the idea is to eventually integrate this helper somehow to lcitool, so
I'm wondering what is it we're trying to solve here. As for the output itself,
if we want to change it, then I'd be probably more inclined towards something
like virt-builder --list. There's massive information redundancy in there -
no need to argue - but because the output is formatted like a simple table,
tools like grep,sort,uniq and cut make the customization of the output for the
average linux user an absolute no-brainer and they always get information they
were looking for easily with almost no added effort.
Erik