Make null param error more informative (#809) #2143
Conversation
| case None => unsafeSetNull(ps, n) | ||
| } | ||
|
|
||
| private def unsafeSetWhatShouldNotBeNull(ps: PreparedStatement, n: Int, a: A, message: Int => String): Unit = |
There was a problem hiding this comment.
I believe this will still allocate the anonymous function (unless JIT is able to elide the allocation?)
There was a problem hiding this comment.
That's actually a good point, but is it so if it's an axisting method passed as an argument? How is it compiled to bytecode?
There was a problem hiding this comment.
https://godbolt.org/z/5sz68x4cE
When I use an anonymous function it seems to extract it into a static method, so there are probably no allocation either way in this case.
I know next to nothing about JVM bytecode though so 😄
There was a problem hiding this comment.
:D
So it seems (but not without a doubt) that it is not affecting performance too much.
Your call then @jatcwang - should we go with it for sake of Developer Experience or is the simpler (already merged) version enough 🙌
There was a problem hiding this comment.
Will keep it simple for now as I feel the merged version already have enough information for people to find the source of null :)
ulfryk
force-pushed
the
fix/uninformative-null-error-fancy
branch
from
e81f52e
to
a15794f
Compare
Same as #2141 but with more information.
The error message string is created only in
if (a == null)branch ( to solve issue spotted here #2141 (comment) )