Preserve comments in config.yml

[samuelallan72]

Is your feature request related to a problem? Please describe.

In testing with Tutor devstacks, I often find myself wanting to temporarily remove an option or add an inline comment to explain something. However, I can't do this, because as soon as Tutor runs, all comments are stripped from the file.

Describe the solution you'd like

It would be great if comments (and bonus: existing order of options) were preserved when Tutor loads and re-saves the config.yml file.

Describe alternatives you've considered

My current workaround is to keep more notes separately, but it's not always ideal, and time is lost copy/pasting things between the config.yml and my notes.

Additional context

N/A

[DawoudSheraz]

Hi.
The config.yml is always re-generated using the config values dictionary (https://github.com/overhangio/tutor/blob/release/tutor/config.py#L271) . It is expected that config.yml won't retain any customization, even if they are comments. It might be possible to add comments on each config value, but I am not sure what value it would yield. This request is the first time (atleast for me) that someone mentioned they add comments in config.

[AntonOfTheWoods]

Hi. The config.yml is always re-generated using the config values dictionary (https://github.com/overhangio/tutor/blob/release/tutor/config.py#L271) . It is expected that config.yml won't retain any customization, even if they are comments. It might be possible to add comments on each config value, but I am not sure what value it would yield. This request is the first time (atleast for me) that someone mentioned they add comments in config.

I definitely +1 this, and would add that the alphabetical reordering on save is also suboptimal. YAML's clear advantage over JSON is that it's much more readable but if you can't group your config items or have comments, it reduces the DX advantage quite a bit. I would also argue that the "promise" of the config.yaml is that we get to control that file, tutor does everything in env. As it stands, tutor controls both.