[cnahed-kpler]

This PR makes sure materializations are loaded before cleanup-related logic runs.
That matters because janitor/cleanup operations depend on materialization definitions already being available.

This occurs here: https://github.com/TobikoData/sqlmesh/blob/6e69ce6cfbceab8e7c5f43cdaa7ac7a5d9015cd6/sqlmesh/core/snapshot/evaluator.py#L2877

[CLAassistant]

All committers have signed the CLA.

[StuffbyYuki]

@cnahed-kpler could you take a look at resolving conflicts?

[StuffbyYuki]

I just kicked off the workflows. Once they all pass, I'll merge!

[StuffbyYuki]

@cnahed-kpler Can you add DCO checks?

[StuffbyYuki]

You can ignore those failed style ci tests, there was a bug from another merge, but the fix got merged.

[StuffbyYuki]

@cnahed-kpler Just following up

[lafirm]

This PR should close this issue #5761 (comment)

[StuffbyYuki]

@cnahed-kpler Bump

[maulanaady]

Reproduced on v0.236.2. sqlmesh janitor fails with "Materialization strategy ... not found" because materializations/ is never imported when the CLI builds the context with load=False. Current workaround: import the custom materialization classes in config.py, would remove our workaround once merged

[lafirm]

Since the PR owner @cnahed-kpler isn’t responding or active, and this change is simple but essential, I can’t resolve the conflicts or prepare it for merge because I don’t have permission to push commits to this branch. @StuffbyYuki Should I open a new PR with the same change and have this one closed?

[StuffbyYuki]

@lafirm that's fine with me. Let me know once you open your PR, I'll close this one

[lafirm]

@StuffbyYuki Created a new PR #6162, so this can be closed now.