On Tue, 16 Dec 2025 22:40:44 GMT, Bradford Wetmore <[email protected]> wrote:
>> Yes, if session is `null` we would allocate the constraints and then fail, >> before the refactoring we failed before allocating constraints. In practice, >> session is never null when we reach that code though. > > Please document assumptions like this. > > This is subtle, and it may take more cycles to maintain as it will likely > have to be rediscovered. > > I'm still not 100% sure the situation is clear in my brain: > `SupportedSignatureAlgorithmConstraints` just disables later checks if the > instance is missing. No functional change here, I just noted that after refactoring if we fail because session is null we would first allocate the constraints object. I'm not assuming that `In practice, session is never null when we reach that code`, just stating the current state of code. Functionality is unchanged with session being null after the refactoring. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/28275#discussion_r2625543160
