Does not possible to use - in usernames, because CLI script replace it to _
vyos@vyos# run show configuration commands | match local-users set service pppoe-server authentication local-users username test@isp-example.com password 'test' [edit] vyos@vyos# sudo cat /run/accel-pppd/pppoe.chap-secrets # username server password acceptable local IP addresses shaper test@isp_example.com * test *
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| In progress | BUG | Dmitry | T3250 PPPoE server: wrong local usernames | ||
| Resolved | jestabro | T3481 Exclude tag node values from key mangling |
On the other hand, it will not be possible to use login with "_"
https://github.com/vyos/vyos-1x/blob/current/data/templates/accel-ppp/chap-secrets.config_dict.tmpl#L6-L8
If we replace all "-"
| replace('_', '-')@jestabro this is due to the key_mangling call for get_config_dict() - I wonder if we can exclude the mangling on the tagNode value itself and how JInja2 behaves with it.
Actually the same applies to us which is here explained for Ansible: https://blog.networktocode.com/post/Exploring-Jinja-Variable-Syntax-in-Ansible/
If we switch to [] notation we could get rid of the key mangler I guess
@c-po this is an excellent idea, and a useful reference for the bracket notation; this will need a bit of design of how best to slice the problem:
I'm looking into the pros/cons of each approach.
With this new information I see little to none reason to keep the key_mangling() workaround. If we manage to transform all nodes into "proper" syntax we can one day drop it.
We shall experiment with the notation on never rewrites - and on success become a design principle.
Also this solves the issue where the user specified a hyphen inside tagNode that also allows underscores - there is yet no way to determine it was a hyphen in the first place.
Yes, I very much like this, and is what I am imaging with 3., above.
As an example, consider any future conf_mode scripts that only use get_config_dict(): in that case it is easy to experiment, as key_mangling is an explicit keyword argument, of default None.
In the case of scripts using, say, get_interface_dict(), key mangling is implicit, and we will need to consider adding an explicit keyword argument, but with the default full mangling (unless we want to correct all current uses now).
Nonetheless, I agree that the first step is to test out the bracket notation; then we can make a decision on how to handle the interim period --- if the interim period involves a few extension to function definitions and use, it will not matter once the key mangling all goes away.