The theme's chrome is worked out once, not per grapheme; still post passes let the GUI rest; a perf gate

somtrmsz's contrast floors made ChromeTheme.fromTheme run its focus-tint
and separator searches (pow calls each), and Output's RowDecoration.styleAt
called it for every grapheme of every highlighted row: the 50k-line file's
render went from 0.11 ms to 13 ms. Pardes.bodyChrome now keeps the theme's
chrome, worked out again only when the theme differs; recolorSyntax asks
for it once a decorated row, and a plain row (nearly every row of a file)
never asks. ReleaseFast, medium fixture, median us, before -> after (main):
render 13150 -> 76 (111), key-down 14658 -> 79 (110), wheel 13589 -> 76
(109), open 14891 -> 2095 (2055), edit-char 16703 -> 3006 (1251; the rest of
that gap is editing and tree-sitter, not this).

Post.animating asked for frames whenever the window had focus and any pass
was ready, so a still pass kept the GUI drawing at the display's rate
(Bloom: 49% of a core idle, in a hidden test window). A pass now says
whether it moves on its own: the CRT (its hum and dither) and a Shadertoy
file whose source reads iTime, iFrame or iDate (shader_build.readsTime; the
flag rides the wire's post message); Bloom, Vignette and Grain are still.
Bloom idle: 49% -> 1.2%, as with no pass.

zig build perf-gate: the 50k-line file's gestures, each's fastest sample
within 3x of the recorded baseline's (test/perf-baseline-<platform>-
<optimize>.json, recorded from this build), run with every unit-test. On
the regressed code it fails at 180x for render.

1 files changed, 15 insertions, 0 deletions

diff --git a/build.zig b/build.zig
index 48987edf..79de8a81 100644
--- a/build.zig
+++ b/build.zig
@@ -1282,6 +1282,18 @@ pub fn build(b: *std.Build) void {
         run_perf.setCwd(b.path("."));
         const perf_step = b.step("perf", "gesture latency or bounded terminal stress (-- [--json] [--reps N] [--base old.json] [--terminal-mib N])");
         perf_step.dependOn(&run_perf.step);
+        // The gate: the 50k-line file's gestures against the baseline this
+        // build shape recorded, each within 3x of its best. Record one with
+        // `zig build perf -Dplatform=P [-Doptimize=O] -- --only medium
+        // --reps 10 --json > test/perf-baseline-P-O.json` (a build that is
+        // known good). ponytail: the numbers are this machine's; another
+        // machine records its own.
+        const perf_gate = b.addRunArtifact(perf);
+        perf_gate.has_side_effects = true;
+        perf_gate.setCwd(b.path("."));
+        perf_gate.addArgs(&.{ "--only", "medium", "--reps", "10", "--base", b.fmt("test/perf-baseline-{t}-{t}.json", .{ platform, optimize }), "--max-ratio", "3" });
+        const perf_gate_step = b.step("perf-gate", "fail when a gesture on the 50k-line file takes over 3x its recorded baseline (test/perf-baseline-<platform>-<optimize>.json)");
+        perf_gate_step.dependOn(&perf_gate.step);
         const perf_tests = b.addTest(.{
             .root_module = b.createModule(.{
                 .target = target,
@@ -1430,6 +1442,9 @@ pub fn build(b: *std.Build) void {
         b.step("history", "compare command output and test/build runtime across jj revisions").dependOn(&run_history.step);
 
         const unit_step = b.step("unit-test", "run native shell and module unit tests");
+        // The perf gate runs with every unit-test (a regression like a
+        // contrast search per grapheme costs 100x, far past the gate).
+        unit_step.dependOn(perf_gate_step);
         const unit_match = testMatch(b, unit_profile, test_filters, unit_step);
         const lspbench_tests = b.addTest(.{ .root_module = lspbench.root_module, .filters = test_filters });
         const run_lspbench_tests = b.addRunArtifact(lspbench_tests);