On Tue Oct 18 10:04:56 2022 +0000, Matteo Bruni wrote:
This should presumably copy data to an offset inside "raw", right? Actually, this poses an issue with preserving the existing contents over the part that's skipped over.
It does.
In the current version of patch 1, I just take the returned pointer and add the offset to it. Writing far enough will cause a C0000005 under win10, so I don't bother checking the returned pointer anymore. (I haven't submitted that version yet because I need to finish fixing the current test failures on it. And I will need to amend the commit to do so.)