• Joined on 2021-11-20
ungleich-gitea commented on issue ungleich-public/cdist#42 2021-11-20 15:30:11 +00:00
__timezone should be singleton

The problem here is, we cannot really offer a transition path to the users. We can warn that it will break and then.. what? I have to continue using non-singleton until upstream switched to…

ungleich-gitea commented on issue ungleich-public/cdist#42 2021-11-20 15:30:10 +00:00
__timezone should be singleton

but can't we allow object_id-s with singletons?

ungleich-gitea commented on issue ungleich-public/cdist#42 2021-11-20 15:30:09 +00:00
__timezone should be singleton

@nico Hm... what if we can support positional parameter for singleton types? It wouldn't be object id but plain positional parameter, optional. It's just that we would need to support it with some…

ungleich-gitea commented on issue ungleich-public/cdist#42 2021-11-20 15:30:07 +00:00
__timezone should be singleton

So there would only be one positional argument allowed and only for singleton types?

ungleich-gitea commented on issue ungleich-public/cdist#42 2021-11-20 15:30:06 +00:00
__timezone should be singleton

I know, it's kind of a hack, but there is already only one positional argument allowed for non-singleton types, and that's object id.

ungleich-gitea commented on issue ungleich-public/cdist#42 2021-11-20 15:30:05 +00:00
__timezone should be singleton

With singleton types you cannot call it object id, since there can be only one singleton type object.

ungleich-gitea commented on issue ungleich-public/cdist#42 2021-11-20 15:30:03 +00:00
__timezone should be singleton

When I think more about it we shouldn't hack singleton type core support. But we should decide on the way of deprecation and replacing the old one with the new one.

ungleich-gitea commented on issue ungleich-public/cdist#42 2021-11-20 15:30:02 +00:00
__timezone should be singleton

What's about to only use it while the non-singleton type is deprecated? So you can map the object id to one parameter. Then, every time the singleton type will be called as a normal object, it will…

ungleich-gitea commented on issue ungleich-public/cdist#42 2021-11-20 15:30:01 +00:00
__timezone should be singleton

@matze But what should the new type be called?

ungleich-gitea commented on issue ungleich-public/cdist#42 2021-11-20 15:29:59 +00:00
__timezone should be singleton

No, I thought that in the transition, you can use the type like this:

__timezone Europe/Berlin
__timezone --timezone Europe/Berlin

With the first call, it's map the "object id" to the…

ungleich-gitea commented on issue ungleich-public/cdist#42 2021-11-20 15:29:58 +00:00
__timezone should be singleton

This would require cdist core code changes, only for this one type. What is deprecated here then? Object id? Or better, type nature?

ungleich-gitea commented on issue ungleich-public/cdist#42 2021-11-20 15:29:57 +00:00
__timezone should be singleton

mentioned in merge request !916

ungleich-gitea commented on issue ungleich-public/cdist#42 2021-11-20 15:29:55 +00:00
__timezone should be singleton

Since no one could really come up with a good solution to the transition issue, do you think we can go forward with !916 and add a "breaking changes" section to the…

ungleich-gitea commented on issue ungleich-public/cdist#42 2021-11-20 15:29:54 +00:00
__timezone should be singleton

closed

ungleich-gitea commented on issue ungleich-public/cdist#41 2021-11-20 15:29:52 +00:00
annoying warnings

@ander Right, the variable name is wrong, it is __target_hostname. I will just remove those warnings. They don't have any special usage meaning anyway.

ungleich-gitea commented on issue ungleich-public/cdist#41 2021-11-20 15:29:51 +00:00
annoying warnings

closed via commit 077989e8fd749339da9a6ca8458b2c73ac71672c

ungleich-gitea commented on issue ungleich-public/cdist#41 2021-11-20 15:29:50 +00:00
annoying warnings

mentioned in commit 077989e8fd749339da9a6ca8458b2c73ac71672c

ungleich-gitea commented on issue ungleich-public/cdist#40 2021-11-20 15:29:48 +00:00
__hosts: modification results in duplicate lines

@ssrq Is the above change: first I used hosts type without alias, run cdist config; then I changed usage of hosts type so that I added alias and run cdist config? If yes, then this cannot be done…

ungleich-gitea commented on issue ungleich-public/cdist#40 2021-11-20 15:29:47 +00:00
__hosts: modification results in duplicate lines

@ssrq Also take a look at the following discussion: https://code.ungleich.ch/ungleich-public/cdist/-/merge_requests/465. I think that Dmitry purposely didn't support alias, multiple names are managed…

ungleich-gitea commented on issue ungleich-public/cdist#40 2021-11-20 15:29:46 +00:00
__hosts: modification results in duplicate lines

changed title from __hosts: {---alias -}modification results in duplicate lines to __hosts: modification results in duplicate lines