Mark Fields comments

Results 56 comments of


                                            Mark Fields

[Prototype] Offline: Add batchId to batch metadata to detect forked containers

Aww way to steal my thunder, microsoft-github-policy-service!! This prototype PR is defunct now, all the behavior has been implemented properly in main 💪

Can't perform horizontal scroll

Just came back to MM after years, and this is tripping me up a lot! My mouse has a horizontal scroll function which is lost in translation on the remote...

Can't perform horizontal scroll

@truong2d is this GH issue list the best place to submit this kind of feedback / feature request? Thanks!

improvement(container-runtime): Further simplifications for dirty/saved logic

Converting to a draft, not expecting to prioritize this before going on Leave. Will finish the tests and pick it up in August.

Removed ScheduleManager and ScheduleManagerCore from container runtime

> The isRuntimeMessage only checks whether the type is MessageType.Operation or not. So, for messages with type like ContainerMessageType.FluidDataStoreOp, it fails Basically, the current code doesn't work for these types...

feat(client): JsonSerializable and JsonDeserialized

It would be a cool follow-up to have JsonString type that is type branded string that implies that if you parse it you'll get JsonDeserialized or something (not sure exactly...

More explicit typing for PendingStateManager.savedOps

Keeping it alive a little longer...

Fix for SharedTree op named idAllocation

> This PR has been automatically marked as stale because it has had no activity for 60 days. It will be closed if no further activity occurs **within 8 days...

Test Infra: Always fail afterEach if expected errors don't match logged errors, even if the test isn't otherwise passing.

Not sure we need AggregateError, both failures will appear without it. After this change, the output will show the core test failure, and then _also_ show a failure under "after...

Test Infra: Always fail afterEach if expected errors don't match logged errors, even if the test isn't otherwise passing.

> I see the intent as to provide more information to the dev, so that relevant failures/errors don't get buried deep. Would just logging those as `warnings` help? I've seen...