Two useful scripts for diff file manipulations.

scripts/split-diff.sh

@@ -0,0 +1,26 @@
+#!/bin/sh
+
+# 1) Splits a patch file into smaller patches - one for each file.
+# 2) Build a directory structure correspoding to paths in patch headers.
+# 3) Copies smaller patches into newly created directories.
+# 4) If a patch introduces a completely new file, it will be put into directory
+#    as a regular file and not diff file.
+
+[ $# == 1 ] || echo "Error: expected single patch file as an argument."
+[ -f $1 ] || echo "Error: no such file '$1'."
+
+path="$1"
+
+splitdiff -a "${path}" >/dev/null
+
+for p in ${path}.part???; do
+  file=`sed -E -n -e '1s/---[[:space:]]+([^[:space:]]+).*/\1/p' ${p}`
+  origsize=`sed -E -n -e '3s/@@ ([^[:space:]]+) .*/\1/p' ${p}`
+  mkdir -p `dirname ${file}`
+  if [ "${origsize}" == "-0,0" ]; then
+    sed -E -n -e '3,$s/^\+(.*)/\1/p' ${p} >${file}
+  else
+    cp ${p} ${file}.diff
+  fi
+  rm ${p}
+done