fix: avoid changing os.environ in Harness #1359
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
It seems like a regression test here isn't required, but let me know if you disagree with that, and I can add one. I'll also do a super-tox run with and without this change - it's surprising to me that it didn't catch it previously, so I'd like to dig into that. |
dimaqq
approved these changes
Sep 5, 2024
james-garner-canonical
approved these changes
Sep 5, 2024
IronCore864
approved these changes
Sep 5, 2024
With main@HEAD I have 111 passes and with this branch 112 (charm-simple-streams is the difference, as expected). So this seems safe. |
tonyandrewmeyer
added a commit
that referenced
this pull request
Sep 5, 2024
Harness needs to provide a `JUJU_VERSION` value to set up the `_JujuContext`, but it doesn't have to be in `os.environ`, it can just be in the dictionary passed to create the `_JujuContext` object. In production, `os.environ` would actually have this, so it's more realistic for it to be present, but it would also have lots of the other `JUJU_` environment variables as well, and we don't want to have Harness simulate all those in the environment - we want people working with the ops tools to access those, not the environment directly. This change [broke the tests of at least one charm](canonical/charm-simple-streams#22) because it patches the environment to have specific values, and then creating the `Harness` object changes that. It seems better for us to not do this - if we did want to populate the environment to mimic Juju then we'd likely want that to be explicit, or done around the event emitting. It was also an accidental backwards compatiblity break.
|
Thanks for this fix folks! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Harness needs to provide a
JUJU_VERSIONvalue to set up the_JujuContext, but it doesn't have to be inos.environ, it can just be in the dictionary passed to create the_JujuContextobject.In production,
os.environwould actually have this, so it's more realistic for it to be present, but it would also have lots of the otherJUJU_environment variables as well, and we don't want to have Harness simulate all those in the environment - we want people working with the ops tools to access those, not the environment directly.This change broke the tests of at least one charm because it patches the environment to have specific values, and then creating the
Harnessobject changes that. It seems better for us to not do this - if we did want to populate the environment to mimic Juju then we'd likely want that to be explicit, or done around the event emitting. It was also an accidental backwards compatiblity break.