From 9b58d2e870fddb5f04e846b0566dd6a613d2f699 Mon Sep 17 00:00:00 2001 From: tv Date: Sun, 8 Nov 2026 18:35:27 +0000 Subject: bench: what screening costs while catching up Review 2, N7, measured first: a node whose document has no policy (a joiner's empty one) applies what it receives unjudged. Screening it too would cost finding every change's paths and judging it. bench_screening takes 1 000, 10 000 and 100 000 changes into an empty document, as one delta brings them: unjudged, as Automerge applies them, against screened (access::screen::apply), with no policy in them (each passes, read at its dependencies, after its paths were found) and with the policy their first change wrote (each judged by the rules). --- docs/development.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'docs') diff --git a/docs/development.md b/docs/development.md index 2fb7ab5..85f6fce 100644 --- a/docs/development.md +++ b/docs/development.md @@ -19,8 +19,8 @@ nor arti; among Mycelium's tests is a swarm simulated on its envelope). The benchmarks are ignored tests named `bench_*`: the document's (writes and sync messages, what signing and checking every change would cost: `bench_change_signatures`, -and knowing every change's paths: `bench_change_paths`; -`document/bench.rs`), the remakes' replay (`remake/bench.rs`), and what +knowing every change's paths: `bench_change_paths`, and judging every change while +catching up: `bench_screening`; `document/bench.rs`), the remakes' replay (`remake/bench.rs`), and what passing changes on costs over the simulated network (`simtest/bench.rs`: messages and bytes per write by kind, propagation delays, for a full mesh, a writer behind NAT (in a full mesh, and in a partial one), two transports, partial meshes of 50 and 100 nodes, lossy links (with delays at random, -- cgit v1.3.1