sops
sops copied to clipboard
getsops
add toml store
Fixes #369.
replace PR #533 as:
- BurntSushi/toml is unmaintained
- I fixed some issues on go-toml regarding key order preservation
Still not perfect, comments are not preserved, but most of it should be here.
this will probably change with the incoming go-toml v2. I've seen no comment for more than 2 months. Does it mean the reviewers are busy or that this has no interest?
No, there’s definitely interest, sorry I missed this. If a new major version of the toml library is going to be released soon, I think it’s best to wait until then before merging this.
no worries, I just wanted to know. I'm participating in this go-toml v2, so, I agree, I'll wait til it's done and then re-update this PR!
I'm keen on this, I could see it being valuable for gitlab-runner configs! :)
It does not seem "go-toml v2" is going to be finalised anytime soon. Last comment here was in eb and now it's almost May. Could we merge this as is and then shift to v2 in a year or when it's finalised? That way we are not blocked and can have toml support.
We are at 2.0.alpha2, guestimating more than a couple months to get to v2 final. I'll make sure sops integration is feasible within the v2 as well, so at least, there is that, there should not be too much time between the release of v2 and the sops toml PR.
If we want to push for this one, we need a review first.
@autrilla can we possibly push for a review sooner? It seems like we're still some while away from go-toml v2. Unless we expect the API to dramatically change, it might be valuable to get TOML support sooner rather than later. We've had to use a couple workarounds to get sops working with TOML and we'd love to see first-class support so that we can iterate and contribute improvements we've written locally
@autrilla I was waiting for the v2 to be out and replace this PR implementation but the document representation of the toml is not in the scope of v2 for now, so it's not really convenient to work for now.
Hence, could this PR be reviewed first?
I was the original author, started it and waited a long time to be merged. by the time it was ready to be merged, go-toml/v2 was released and the author did not want to have the internal struct public, so it's more difficult to apply the same strategy I used with the v1, hence i stalled :/