On Tue, 29 Nov 2022 07:39:59 GMT, Andrey Turbanov <aturba...@openjdk.org> wrote:

> IDEA now highlights unnecessary `rem.offset=1` assignments, which are above 
> modified code. Now we always overwrite it with `0` anyway.
> 
> https://github.com/openjdk/jdk/blob/6f06f440bcf8a5db379b80e8765af38a15449356/src/java.base/share/classes/java/math/MutableBigInteger.java#L1498
> 
> https://github.com/openjdk/jdk/blob/6f06f440bcf8a5db379b80e8765af38a15449356/src/java.base/share/classes/java/math/MutableBigInteger.java#L1504
> 
> https://github.com/openjdk/jdk/blob/6f06f440bcf8a5db379b80e8765af38a15449356/src/java.base/share/classes/java/math/MutableBigInteger.java#L1520
> 
> Perhaps we can drop them too.

Hmm. Looks to me like the earlier assignments of 1 to the rem object's offset 
field are to preserve inter-field semantic constraints. So while it is true 
this field is not used before it is overwritten circa line 1536, I'd reluctant 
to "break" the object constraints in the interim since it could complicated 
future maintenance, event with commenting.

Thoughts @bplb or @rgiulietti ?

-------------

PR: https://git.openjdk.org/jdk/pull/11395

Reply via email to