On Tue, 26 Mar 2024 07:44:22 GMT, Magnus Ihse Bursie <i...@openjdk.org> wrote:
> > I have a concern since the null check bailout involves
> > THROW_NULL_PDATA_IF_NOT_DESTROYED, which is no longer accurate if we remove
> > the pData local.
>
> The name of the macro is not great, but it does not involve pData (the bad
> NPE error message notwithstanding):
>
> ```
> #define THROW_NULL_PDATA_IF_NOT_DESTROYED(peer) { \
> jboolean destroyed = JNI_GET_DESTROYED(peer); \
> if (destroyed != JNI_TRUE) { \
> env->ExceptionClear(); \
> JNU_ThrowNullPointerException(env, "null pData"); \
> } \
> }
> ```
>
> So you can go ahead and replace the pData references with the variable that
> will eventually be used.
Alright, will do. Maybe as a further improvement, I can inline
THROW_NULL_PDATA_IF_NOT_DESTROYED at its callsites and replace the bad
NullPointerException error message with the proper null pointer name. Since
Phil isn't here, what do you think?
Regardless, I really hope I can get this in by Thursday. University for me
officially ramps up into _very_ high gear about that time, and I doubt I can
juggle both JDK work and it all at once by then
-------------
PR Comment: https://git.openjdk.org/jdk/pull/15096#issuecomment-2019839348
