Piotr Findeisen

do we want to run CI on maintenance branches such as 1.5.x?

My personal preference is with this one. in #10514 I didn't like the fact that i foresaw 2.x release branches and i didn't foresee 3.x release branches (or 4.x, ...)....

thank you @nastra for your feedback i actually think we can do even simpler: remove any branch patterns. it won't make the ci run on docs or gh-pages branches --...

changed the PR with the above in mind. it's now simpler. but will also undraft https://github.com/apache/iceberg/pull/10514. also sent https://github.com/apache/iceberg/pull/10578 to fix current 1.5.x branch

https://github.com/apache/iceberg/pull/10514 merged, so closing this one

There is a regression for CREATE TABLE https://github.com/apache/datafusion/pull/12864#issuecomment-2437123342 This should be resolved before release (or judged OK to break).

> The syntax with both sessions and SQL functions: > > ```sql > WITH > SESSION a = 1 > FUNCTION x(a BIGINT) RETURNS BIGINT RETURN 1 > SELECT x()...

> A recent [supply chain attack](https://arstechnica.com/information-technology/2025/03/supply-chain-attack-exposing-credentials-affects-23k-users-of-tj-actions/) has made it extremely apparent that github workflows should only use actions that are tied to a specific hash, not a version. This applies...

> "casting should not be changed after planning" if i said exactly that I should stand corrected. _coercions_ is somewhat that should be applied during analysis/initial planning phase. Coercion rules...

It's very natural to think about file-level vs table level as same thing as SQL coercions, but there is an important distinction. SQL has its own semantics and table provider...