> What we need next is (a) a clean patch, and (b) thorough testing that
> the standard handles are indeed not inherited by sub-processes of
> Emacs when the patch is applied.
What should be the testing for (b)?
One thing I noticed when I used a version with broken stdout is that the async
compilation (native or via async package) doesn't work properly. Those errors
don't show up anymore for the patch of using SetHandleInformation.
I have been using the UCRT version for a while and notice nothing has gone
wrong so far even with the `native-comp-jit-compilation'. It would be great if
someone can share how I can do further testing on this.
----
Kien