Navin Agarwal

Results 23 issues of Navin Agarwal

## Description Writing the GC data at root of the summary was enabled via feature flags since version 0.56 - https://github.com/microsoft/FluidFramework/commit/7a62ce100db0064c67d7626bd47846a2817ffe7e. It was enabled by default in version 1.1 -...

area: runtime
area: tests
public api change
base: next

This allows tests using testObjectProvider to create custom data objects and runtime factories with the current compat versions. For example, if the tests are running in version N-1, it will...

area: tests
public api change
base: main

## Background This change is aimed at removing the tracking of basePath, localPath and additionalPath from SummarizerNode. These are used in case a node did not change and it's summary...

area: runtime
public api change
base: main

## Background This change is aimed at removing the tracking of basePath, localPath and additionalPath from SummarizerNode. These are used in case a node did not change and it's summary...

area: runtime
public api change
base: main

https://github.com/microsoft/FluidFramework/pull/16885 added a new retry mechanism in case of summarization failures. This was rolled out with A/B testing and the experiment was successful. This PR makes the new retry mechanism...

documentation
area: runtime
area: tests
public api change
base: main

Added a test that benchmarks the `initializeOrUpdateGCState` method in case of high numbers of unreferenced nodes. [AB#7649](https://dev.azure.com/fluidframework/235294da-091d-4c29-84fc-cdfc3d90890b/_workitems/edit/7649) **Todo** - [ ] Figure out how to actually benchmark. Currently, the test...

area: runtime
base: main

GC is enabled by default and has been this way for quite a while. Specifically, the mark phase which marks unreferenced objects as such in the summary is always enabled....

area: runtime
area: tests
public api change
base: main

The error thrown during summarization go through multiple layers - container runtime -> summary generator -> running summarizer. Currently, the error type is `Error` and it gets lost while is...

area: runtime
public api change
base: main

Currently, GC runs will fullGC a little often which also results in running summarize with fullTree which is expensive. Here is the list of scenarios where full GC runs today...

area: runtime
public api change
base: main

This is a pre-cursor to https://github.com/microsoft/FluidFramework/pull/21186 which removes `trackState`. In `SummarizerNode::summarize`, removed the validation that a summary in progress. This validation exists to ensure that `startSummary` was called before `summarize`...

area: framework
area: runtime
area: tests
base: main