Conversation
- Since this section is going to be long due to the need of documenting the different backends then giving it its own section seems reasonable. Additionally, by introducing fetchers before derivations and after the basics of the language, it guides the lector through the mental model that in Nix you build your derivations from the data you fetch in a reproducible way in the first place - Start by documenting the `git` backend and then I'll be adding more
| @@ -0,0 +1,147 @@ | |||
| **Warning**: | |||
| This section is **experimental** and its interface is subject to change. | |||
There was a problem hiding this comment.
How can this be experimental when all the old fetchers are now using this under the hood?
There was a problem hiding this comment.
I added the warning because of this small detail:
nix/src/libexpr/primops/fetchTree.cc
Line 179 in bceda30
There was a problem hiding this comment.
How can this be experimental when all the old fetchers are now using this under the hood?
I don’t think here’s a contradiction here: This is indeed used internally, but the interface isn’t stable yet. So it makes sense to mark this as experimental
|
The fetchers stuff is already mostly documented in the |
I agree, let's not duplicate everything, but what about the other way around? |
|
I marked this as stale due to inactivity. → More info |
5 participants
due to the need of documenting the different backends
then giving it its own section seems reasonable.
Additionally, by introducing fetchers before derivations and
after the basics of the language,
it guides the lector through the mental model
that in Nix you build your derivations from the data you fetch
in a reproducible way in the first place
gitbackend and then I'll be adding moreMentions #5463