Hi,

I read on Npgsql 9.0 release notes, that they are considering dropping support for synchronous APIs. Link: https://www.npgsql.org/doc/release-notes/9.0.html

I want to ask if LLBLGen will be ready to work without those APIs, when Npgsql 10 comes out?

I replied last year in their issue: https://github.com/npgsql/npgsql/issues/5865

We have an async api, so it should be fine in general, IF you only call async methods. However we don't implement CommitAsync calls as that's not in .netstandard2.0. We ship an older npgsql for the driver so it's only at runtime.

My personal opinion is that this is a stupid move of them. Removing APIs that are used by many is beyond pointless. You can deprecate things and signal to the user 'hey it's better to use the alternative' but there's little else to it.

But in the end it's all what-if, they haven't released v10 yet and in v10 it still should work ok. We'll see.

Thanks for your reply.

Of course, we are also not happy if those sync methods stopped worked, primarily because we have code that is not using async calls, that would have to be rewritten.

You are right that Npgsql 10 should work (if they add a toggle to enable sync I/O), so it probably will be in Npgsql 11 that it could be a real problem.

In any case, if we ever come to a situation where we can't use the latest bits from Microsoft, and LLBLGen is preventing us to upgrade, then that would probably mean we would have to look for alternatives. So the goal with my post was merely to give me some kind of hope that we can use LLBLGen for many years forward, even if Microsoft does things we disagree with :-)

mbp wrote:

Thanks for your reply.

Of course, we are also not happy if those sync methods stopped worked, primarily because we have code that is not using async calls, that would have to be rewritten.

You are right that Npgsql 10 should work (if they add a toggle to enable sync I/O), so it probably will be in Npgsql 11 that it could be a real problem.

In any case, if we ever come to a situation where we can't use the latest bits from Microsoft, and LLBLGen is preventing us to upgrade, then that would probably mean we would have to look for alternatives. So the goal with my post was merely to give me some kind of hope that we can use LLBLGen for many years forward, even if Microsoft does things we disagree with :-)

If the current npgsql does what you need, you don't need to upgrade to newer version that ditches existing api's. So I guess it's only for new code targeting .net 10+ that might be affected. Existing projects targeting .net 8 with e.g. npgsql 8/9 will keep on working just fine.

I don't want to give you false hope, as in the end it's not a problem we create, so there's little we can do but wait and see. But for the coming years (2 at least I think) I wouldn't worry about it