Replies: 1 comment
-
A couple of notes:
Multi-output recipes are just weird 😬 |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Question
I am currently trying to figure out how to backport this change with a repodata patch:
Change: https://github.com/conda-forge/seaborn-feedstock/pull/35/files#diff-f3725a55bf339595bf865fec73bda8ac99f283b0810c205442021f29c06eea9aR40-R41
In
index.json
an exact pinning with this mechanism seems to show up as<package-name> <verision> <build>
in thedepends
section, so my naive guess for a patch was:However,
${build}
does not seem to be exposed in the patching yaml 🤔 https://github.com/conda-forge/conda-forge-repodata-patches-feedstock/blob/main/recipe/patch_yaml_utils.py#L10-L19Is there currently a way to backport the exact pin, and if not, would it make sense to expose
${build}
in the yaml file?Exposing build variable and applying patch
After exposing
build
, the repodata patch would do the following:This assumes that the
build
variable is the same for the two outputsseaborn
andseaborn-base
. Empirically this seems to hold, but since I don't know how the build string is generated exactly it probably would be good to sanity-check this.For comparison, this is how the index looks like for
pin_subpackage("seaborn-base", exact=True)
whereseaborn
pinsseaborn-base
in its dependencies.Beta Was this translation helpful? Give feedback.
All reactions