Faster multi-element serialization for primitive types#1872
Conversation
AArnott
left a comment
There was a problem hiding this comment.
Most of the hardware accelerated code is beyond me. But I have a couple questions at least.
AArnott
left a comment
There was a problem hiding this comment.
The product changes look good. How are we looking for test coverage? In particular, given the 16-element boundary that is interesting, it would seem prudent to have tests that cover lists of various sizes including at and around that 16-length boundary. Do we already have such tests, and are they running on .NET 8 such that they'll cover your new code?
|
There is an awful amount of noise in the failure logs, but there does appear to be a legitimate failure here. |
|
I've fixed bugs. I have a question about |
Given the length of the existing file, my vote is create a second file. |
|
Do you want me to merge this as-is, or do you want to add the deserialization optimization to it first? |

Rewrite of #1744