Fix rendering of Column objects in dialect kwargs like postgresql_inc…

[Ajaysingh-2003]

Fixes #1258

This PR updates _render_dialect_kwargs_items in alembic/autogenerate/render.py to properly handle list/tuple values containing Column objects, such as postgresql_include.

The implementation mirrors the logic used in _get_index_rendered_expressions, ensuring consistent rendering behavior.

Changes:

• Proper handling of Column objects inside dialect kwargs
• Fix invalid Python code generation in migrations

Tests:

• Added tests covering both create_index and drop_index cases

This ensures Alembic generates valid migration scripts when using postgresql_include.

[CaselIT]

thanks. looks almost ok:

• left a comment
• could you add a changenote here https://github.com/sqlalchemy/alembic/tree/main/docs/build/unreleased? the number is the issue. Feel free to add Pull request courtesy of <your> at the end

[CaselIT]

let's try a tuple here

[Ajaysingh-2003]

Updated, both tests now use tuples.

[CaselIT]

I suggested only one to that we kept leveraging both code paths in the test. if it's not too much of a both I would prefer if one of the two test kept using list, so that ensure that both work.

thanks!

[Ajaysingh-2003]

Done! I updated test_render_drop_index to use a Tuple as requested, while keeping test_render_add_index as a List to comprehensively test both code paths. The tests compile properly.

[Ajaysingh-2003]

thanks. looks almost ok:

• left a comment
• could you add a changenote here https://github.com/sqlalchemy/alembic/tree/main/docs/build/unreleased? the number is the issue. Feel free to add Pull request courtesy of <your> at the end

Done — added the changenote in docs/build/unreleased/1258.rst and switched the tests to use tuples.

[CaselIT]

the fails seem unrelated. thanks!

[sqla-tester]

OK, this is sqla-tester setting up my work on behalf of CaselIT to try to get revision 2a0441c of this pull request into gerrit so we can run tests and reviews and stuff

[sqla-tester]

New Gerrit review created for change 2a0441c: https://gerrit.sqlalchemy.org/c/sqlalchemy/alembic/+/6647

[Ajaysingh-2003]

the fails seem unrelated. thanks!

heyy. Any updates on the PR?? Anything to change or update?

[CaselIT]

Sorry I was mostly waiting for Mike to get back from holiday

[Ajaysingh-2003]

Sorry I was mostly waiting for Mike to get back from holiday

Alright, thanks.

[sqla-tester]

Michael Bayer (zzzeek) wrote:

I moved the logic added to _render_dialect_kwargs_items directly into _render_potential_expr() so it can take effect in general.

@CaselIT take another look if you can

View this in Gerrit at https://gerrit.sqlalchemy.org/c/sqlalchemy/alembic/+/6647

[sqla-tester]

Federico Caselli (CaselIT) wrote:

makes sense, left a question, but otherwise seems good

View this in Gerrit at https://gerrit.sqlalchemy.org/c/sqlalchemy/alembic/+/6647

• alembic/autogenerate/render.py (line 635): do we actually need this now? can't we use the if below for ClauseElement?

[sqla-tester]

Michael Bayer (zzzeek) wrote:

code review left on gerrit

View this in Gerrit at https://gerrit.sqlalchemy.org/c/sqlalchemy/alembic/+/6647

• alembic/autogenerate/render.py (line 635): it's still two isinstance() checks either way, so i just did it this way. how would you do it?

[sqla-tester]

Federico Caselli (CaselIT) wrote:

code review left on gerrit

View this in Gerrit at https://gerrit.sqlalchemy.org/c/sqlalchemy/alembic/+/6647

• alembic/autogenerate/render.py (line 635): we did not need this before, that's all I'm saying. but it's fine to keep, no other test is affected so that's fine