Closed aalowlevel closed 1 year ago
Thanks for feedback. The problem is we added close
support and that requires call
returning error.
You might consider a callback for
Well, I don't know. In this way of design, the close
function as well.call
function forces the app error type to be wrapped into the error of its type .
I'll try to figure out a way.
@programatik29 we can provide two interfaces since clearly both have their pros and cons? Maybe hide the close
method behind a feature flag? What do you think?
@programatik29 we can provide two interfaces since clearly both have their pros and cons? Maybe hide the
close
method behind a feature flag? What do you think?
Maybe we can just add a panicking call method and solve this. If close is never called, it would never panic. With proper documentation I think this might be a good solution.
Sure, any specific name you have in mind?
Sure, any specific name you have in mind?
@Czocher Naming it call_unwrap
and documenting panic behavior should be fine.
Hello. It's always problematic when i want to write async wrappers and i am not interested to work with neither
tokio_rusqlite
type errors norrusqlite
type errors:I just want to work like i used to in
0.3.0
. Just go back to the old style.