Andrea Canciani

@roji I implemented the SqlExpressionFactory changes in https://github.com/dotnet/efcore/pull/34002 and based this PR on top of that (in the first PR I update the test so that they keep checking the...

The first commit belongs to #34078 (I rebased on top of that to have a working build). Apart from that, it is ready for review 🚀

Rebased to resolve merge conflict

Some specific `bool?` expression already work (example: propagating the value of a nullable column). The problem is related to the conversion from Boolean to `BIT` (it mixes false and `NULL`...

I am investigating the possible translations. Currently the best candidate for an expression `x` seems to be: ```sql CASE WHEN x IS NULL THEN NULL WHEN x THEN CAST(1 AS...

As of 7e313ab6cff18ef5c58bec565fa7da7ca6bba36f XOR is supported. I can try and rebase https://github.com/dotnet/efcore/compare/main...roji:efcore:MoreBitwise to also tackle left and right shift if it makes sense, but I would need some guidance around...

This does not tackle the issue for `bool?` (handled in https://github.com/dotnet/efcore/pull/33940) This relies on the improvements to `COALESCE` optimization (specifically https://github.com/dotnet/efcore/pull/33938) to avoid regressions.

Rebased to fix merge conflicts and to make commits more readable now that #33938 is merged.

The PostgreSQL documentation provides some words of warning around side effects: - regarding the [evaluation order](https://www.postgresql.org/docs/current/sql-expressions.html#SYNTAX-EXPRESS-EVAL) - and how they are [modeled/used in functions](https://www.postgresql.org/docs/current/xfunc-volatility.html)

@maumar IIUC in the case you describe the analysis is returning correct (inaccurate) nullability information; instead of reporting the more precise result `(value: constant C, nullable: false)` it returns `(value:...