The upstreaming goal from #2: contribute the sampling engine back to microsoft/mimalloc, where the maintainer signaled receptiveness (microsoft/mimalloc#1070) and the Datadog draft (microsoft/mimalloc#1266) stalled. Mechanics are settled in #2: cut an ephemeral pr/* branch from upstream/main, cherry-pick C-only commits (the never-mix commit discipline exists for exactly this), push to a throwaway fork, open the upstream PR referencing #1070/#1266 and crediting the Datadog design. Scope decision to make first: offer the minimal sampling engine + hooks (closest to #1266, easiest to land) vs including the M2 query/snapshot API (bigger surface, harder review). Suggest minimal-first. Wait for #23 to land so the offered engine is final.
🤖 Generated with Claude Code
The upstreaming goal from #2: contribute the sampling engine back to microsoft/mimalloc, where the maintainer signaled receptiveness (microsoft/mimalloc#1070) and the Datadog draft (microsoft/mimalloc#1266) stalled. Mechanics are settled in #2: cut an ephemeral
pr/*branch fromupstream/main, cherry-pick C-only commits (the never-mix commit discipline exists for exactly this), push to a throwaway fork, open the upstream PR referencing #1070/#1266 and crediting the Datadog design. Scope decision to make first: offer the minimal sampling engine + hooks (closest to #1266, easiest to land) vs including the M2 query/snapshot API (bigger surface, harder review). Suggest minimal-first. Wait for #23 to land so the offered engine is final.🤖 Generated with Claude Code