Fix error in clip reduction

One bit of logic from the GLSL (piet-gpu) version of clip_reduce didn't get copied over.

Fixes #648

vello_shaders/shader/clip_reduce.wgsl

diff --git a/vello_shaders/shader/clip_reduce.wgsl b/vello_shaders/shader/clip_reduce.wgsl
index bdbc0e9..a181201 100644
--- a/vello_shaders/shader/clip_reduce.wgsl
+++ b/vello_shaders/shader/clip_reduce.wgsl
@@ -47,6 +47,9 @@
     workgroupBarrier();
     let size = sh_bic[0].b;
     bic = Bic();
+    if local_id.x + 1u < WG_SIZE {
+        bic = sh_bic[local_id.x + 1u];
+    }
     if is_push && bic.a == 0u {
         let local_ix = size - bic.b - 1u;
         sh_parent[local_ix] = local_id.x;

vello_shaders/src/cpu/clip_reduce.rs

diff --git a/vello_shaders/src/cpu/clip_reduce.rs b/vello_shaders/src/cpu/clip_reduce.rs
index b1c2d21..71ad912 100644
--- a/vello_shaders/src/cpu/clip_reduce.rs
+++ b/vello_shaders/src/cpu/clip_reduce.rs
@@ -24,10 +24,10 @@
             let inp = clip_inp[global_ix].path_ix;
             let is_push = inp >= 0;
             let bic = ClipBic::new(1 - is_push as u32, is_push as u32);
-            bic_reduced = bic.combine(bic_reduced);
             if is_push && bic_reduced.a == 0 {
                 scratch.push(global_ix as u32);
             }
+            bic_reduced = bic.combine(bic_reduced);
         }
         reduced[wg_ix as usize] = bic_reduced;
         for (i, parent_ix) in scratch.iter().rev().enumerate() {