ci: reduce to 2 attempts, nuke build dir on retry

Replace 3-attempt loops with 2 attempts everywhere. On the single
retry, nuke the entire build directory rather than partial cleanup.

Applies to retry_build() in retry-build.sh, nick-fields/retry in
test.yml (Build and Pre-Build steps), and bench.yml Setup & Build.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: sbryngelson

.github/scripts/retry-build.sh

@@ -1,29 +1,13 @@
 #!/bin/bash
-# Provides retry_build(): 3-attempt loop with configurable cleanup.
-# Set RETRY_CLEAN_CMD to override cleanup (default: rm -rf build/staging build/install build/lock.yaml).
+# Provides retry_build(): 2-attempt loop.
+# On failure of attempt 1, nukes the entire build directory before attempt 2.
 # Set RETRY_VALIDATE_CMD to run a post-build validation; failure triggers a retry.
 # Usage: source .github/scripts/retry-build.sh
 #        retry_build ./mfc.sh build -j 8 --gpu acc
 
-_retry_clean() {
-    local clean_cmd="$1"
-    local attempt="${2:-1}"
-    if [ "$attempt" -ge 2 ]; then
-        echo "  Attempt $attempt failed — nuking entire build directory for a clean retry."
-        rm -rf build 2>/dev/null || true
-        return 0
-    fi
-    if eval "$clean_cmd" 2>/dev/null; then
-        return 0
-    fi
-    echo "  Cleanup failed; falling back to best-effort rm."
-    rm -rf build/staging build/install build/lock.yaml 2>/dev/null || true
-}
-
 retry_build() {
-    local clean_cmd="${RETRY_CLEAN_CMD:-rm -rf build/staging build/install build/lock.yaml}"
     local validate_cmd="${RETRY_VALIDATE_CMD:-}"
-    local max_attempts=3
+    local max_attempts=2
     local attempt=1
     while [ $attempt -le $max_attempts ]; do
         echo "Build attempt $attempt of $max_attempts..."
@@ -32,7 +16,8 @@ retry_build() {
                 if ! eval "$validate_cmd"; then
                     echo "Post-build validation failed on attempt $attempt."
                     if [ $attempt -lt $max_attempts ]; then
-                        _retry_clean "$clean_cmd" "$attempt"
+                        echo "  Nuking build directory before retry..."
+                        rm -rf build 2>/dev/null || true
                         sleep 5
                         attempt=$((attempt + 1))
                         continue
@@ -46,7 +31,8 @@ retry_build() {
             return 0
         fi
         if [ $attempt -lt $max_attempts ]; then
-            _retry_clean "$clean_cmd" "$attempt"
+            echo "  Build failed — nuking build directory before retry..."
+            rm -rf build 2>/dev/null || true
             sleep 30
         else
             echo "Build failed after $max_attempts attempts."

.github/workflows/bench.yml

@@ -106,7 +106,7 @@ jobs:
         if: matrix.build_script != ''
         uses: nick-fields/retry@v3
         with:
-          max_attempts: 3
+          max_attempts: 2
           retry_wait_seconds: 60
           timeout_minutes: 150
           command: |
@@ -118,9 +118,7 @@ jobs:
             wait $pid2; e2=$?
             [ $e1 -eq 0 ] && [ $e2 -eq 0 ]
           on_retry_command: |
-            (cd pr     && ./mfc.sh clean) &
-            (cd master && ./mfc.sh clean) &
-            wait
+            rm -rf pr/build master/build
 
       - name: Bench (Master v. PR)
         run: bash pr/.github/scripts/run_parallel_benchmarks.sh ${{ matrix.device }} ${{ matrix.interface }} ${{ matrix.cluster }}

.github/workflows/test.yml

@@ -249,11 +249,11 @@ jobs:
         if:   matrix.cluster != 'phoenix'
         uses: nick-fields/retry@ce71cc2ab81d554ebbe88c79ab5975992d79ba08 # v3
         with:
-          max_attempts: 3
+          max_attempts: 2
           retry_wait_seconds: 60
           timeout_minutes: 60
           command: bash .github/workflows/${{ matrix.cluster }}/build.sh ${{ matrix.device }} ${{ matrix.interface }}
-          on_retry_command: ./mfc.sh clean
+          on_retry_command: rm -rf build
 
       - name: Test
         run:  bash .github/workflows/${{ matrix.cluster }}/submit.sh .github/workflows/${{ matrix.cluster }}/test.sh ${{ matrix.device }} ${{ matrix.interface }} ${{ matrix.shard }}
@@ -340,11 +340,11 @@ jobs:
         if:   matrix.cluster != 'phoenix'
         uses: nick-fields/retry@ce71cc2ab81d554ebbe88c79ab5975992d79ba08 # v3
         with:
-          max_attempts: 3
+          max_attempts: 2
           retry_wait_seconds: 60
           timeout_minutes: 120
           command: bash .github/scripts/prebuild-case-optimization.sh ${{ matrix.cluster }} ${{ matrix.device }} ${{ matrix.interface }}
-          on_retry_command: ./mfc.sh clean
+          on_retry_command: rm -rf build
 
       - name: Run Case-Optimization Tests
         run:  bash .github/workflows/${{ matrix.cluster }}/submit.sh .github/scripts/run_case_optimization.sh ${{ matrix.device }} ${{ matrix.interface }}