validator_ragel: Add ncval tests from the old validator

src/trusted/validator_ragel/unreviewed/parse_hex.py

Index: src/trusted/validator_ragel/unreviewed/parse_hex.py
diff --git a/src/trusted/validator_ragel/unreviewed/parse_hex.py b/src/trusted/validator_ragel/unreviewed/parse_hex.py
new file mode 100644
index 0000000000000000000000000000000000000000..9bb90a366197f2e7e72e2da82729f30536f69f2b
--- /dev/null
+++ b/src/trusted/validator_ragel/unreviewed/parse_hex.py
@@ -0,0 +1,84 @@
+#!/usr/bin/python
+# Copyright (c) 2012 The Native Client Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+"""Converts outputs from the old validator into a list of general errors.
+
+The outputs then would be compared for exact match with what the
+script run_ncval_tests.py outputs for the test.
+
+Works under assumption that the corresponding input files (*.hex) maintain a
+separation of instructions: one instruction per line in .hex file.  We could get
+the same information from the validator, but, luckily, we do not have to do
+that.
+"""
+
+import re
+import sys
+
+import run_ncval_tests
+
+
+def WriteFile(filename, data):
+  fh = open(filename, "w")
+  try:
+    fh.write(data)
+  finally:
+    fh.close()
+
+
+def Main(argv):
+  argv = argv[1:]
+  offset = None
+  for arg in argv:
+    assert(arg[-5:] == '.rval')
+
+    # Parse the golden file 1st time, find bundle_crosses.
+    bundle_crosses = []
+    for line in open(arg, 'r').readlines():
+      bundle_error_m = re.match(
+          r'VALIDATOR: ERROR: ([0-9a-zA-Z]+): Bad basic block alignment', line)
+      if bundle_error_m:
+        bundle_crosses.append(int(bundle_error_m.group(1), 16))
+
+    # Find offsets for all instructions that cross a bundle.
+    err_offsets = {}
+    insts = run_ncval_tests.InstByteSequence()
+    insts.Parse(arg[:-5] + '.hex')
+    for bundle_offset in set(bundle_crosses):
+      off = bundle_offset - 1
+      while off >= 0:
+        if insts.HasOffset(off):
+          err_offsets[off] = ['crosses boundary']
+          break
+        off -= 1
+
+    # Parse the golden file 2nd time.  Find other offsets that cause errors.  An
+    # error report takes 2 sequential lines.
+    seen_offset = False
+    for line in open(arg, 'r').readlines():
+      ln = line.rstrip()
+      off_m = re.match(r'VALIDATOR: ([0-9a-z]+):', ln)
+      if off_m:
+        seen_offset = True
+        prev_offset = offset
+        offset = int(off_m.group(1), 16)
+        continue
+      val_error_m = re.match(r'VALIDATOR: ERROR:', ln)
+      if val_error_m and seen_offset and prev_offset != offset:
+        err_offsets.setdefault(offset, []).append('validation error')
+      seen_offset = False
+
+    # Output the error messages in offset order.
+    golden_text = ''
+    for off, msg_lst in sorted(err_offsets.iteritems()):
+      for msg in msg_lst:
+        golden_text += 'offset 0x%x: %s\n' % (off, msg)
+    filename = arg[:-5] + '.val.ref'
+    print 'writing file: %s' % filename
+    WriteFile(filename, golden_text)
+
+
+if __name__ == '__main__':
+  sys.exit(Main(sys.argv))