Add option to force filetype=asm for testing

pydir/run-pnacl-sz.py

 #!/usr/bin/env python2
 
 import argparse
 import itertools
 import os
 import re
 import subprocess
 import sys
 import tempfile
 
@@ skipping 22 matching lines @@
     """Run the pnacl-sz compiler on an llvm file.
 
     Takes an llvm input file, freezes it into a pexe file, converts
     it to a Subzero program, and finally compiles it.
     """
     argparser = argparse.ArgumentParser(
         description='    ' + main.__doc__,
         formatter_class=argparse.ArgumentDefaultsHelpFormatter)
     argparser.add_argument('--input', '-i', required=True,
                            help='LLVM source file to compile')
+    argparser.add_argument('--output', '-o', required=False,
+                           help='Output file to write')
     argparser.add_argument('--insts', required=False,
                            action='store_true',
                            help='Stop after translating to ' +
                            'Subzero instructions')
     argparser.add_argument('--no-local-syms', required=False,
                            action='store_true',
                            help="Don't keep local symbols in the pexe file")
     argparser.add_argument('--llvm', required=False,
                            action='store_true',
                            help='Parse pexe into llvm IR first, then ' +
@@ skipping 17 matching lines @@
                            action='store_true',
                            help='Assemble the output')
     argparser.add_argument('--disassemble', required=False,
                            action='store_true',
                            help='Disassemble the assembled output')
     argparser.add_argument('--dis-flags', required=False,
                            action='append', default=[],
                            help='Add a disassembler flag')
     argparser.add_argument('--filetype', default='iasm', dest='filetype',
                            choices=['obj', 'asm', 'iasm'],
-                           help='Output file type.  Default %(default)s.')
+                           help='Output file type.  Default %(default)s')
+    argparser.add_argument('--forceasm', required=False, action='store_true',
+                           help='Force --filetype=asm')
     argparser.add_argument('--target', default='x8632', dest='target',
                            choices=['x8632','arm32','mips32'],
-                           help='Target architecture.  Default %(default)s.')
+                           help='Target architecture.  Default %(default)s')
     argparser.add_argument('--echo-cmd', required=False,
                            action='store_true',
                            help='Trace command that generates ICE instructions')
     argparser.add_argument('--tbc', required=False, action='store_true',
                            help='Input is textual bitcode (not .ll)')
     argparser.add_argument('--expect-fail', required=False, action='store_true',
                            help='Negate success of run by using LLVM not')
     argparser.add_argument('--args', '-a', nargs=argparse.REMAINDER,
                            default=[],
                            help='Remaining arguments are passed to pnacl-sz')
     argparser.add_argument('--sandbox', required=False, action='store_true',
-                           help='Sandboxes the generated code.')
+                           help='Sandboxes the generated code')
 
     args = argparser.parse_args()
     pnacl_bin_path = args.pnacl_bin_path
     llfile = args.input
 
     if args.llvm and args.llvm_source:
       raise RuntimeError("Can't specify both '--llvm' and '--llvm-source'")
 
     if args.llvm_source and args.no_local_syms:
       raise RuntimeError("Can't specify both '--llvm-source' and " +
                          "'--no-local-syms'")
 
     if args.llvm_source and args.tbc:
       raise RuntimeError("Can't specify both '--tbc' and '--llvm-source'")
 
     if args.llvm and args.tbc:
       raise RuntimeError("Can't specify both '--tbc' and '--llvm'")
 
+    if args.forceasm:
+      if args.filetype == 'asm':
+        pass
+      elif args.filetype == 'iasm':
+        # TODO(sehr) implement forceasm for iasm.
+        pass
+      elif args.filetype == 'obj':
+        args.filetype = 'asm'
+        args.assemble = True
+
     cmd = []
     if args.tbc:
       cmd = [os.path.join(pnacl_bin_path, 'pnacl-bcfuzz'), llfile,
              '-bitcode-as-text', '-output', '-', '|']
     elif not args.llvm_source:
       cmd = [os.path.join(pnacl_bin_path, 'llvm-as'), llfile, '-o', '-', '|',
              os.path.join(pnacl_bin_path, 'pnacl-freeze')]
       if not args.no_local_syms:
         cmd += ['--allow-local-symbol-tables']
       cmd += ['|']
@@ skipping 10 matching lines @@
       cmd += ['-verbose', 'inst', '-notranslate', '-threads=0']
     if not args.llvm_source:
       cmd += ['--bitcode-format=pnacl']
       if not args.no_local_syms:
         cmd += ['--allow-local-symbol-tables']
     if args.llvm or args.llvm_source:
       cmd += ['--build-on-read=0']
     else:
       cmd += ['--build-on-read=1']
     cmd += ['--filetype=' + args.filetype]
+    # Redirecting the output file needs to be done through the script because

[Jim Stichnoth, 2016/01/13 00:36:44]

This comment should probably be moved closer to the "-o" block.

[sehr, 2016/01/13 18:15:35]

Done.

+    # forceasm may introduce a new temporary file between pnacl-sz and llvm-mc.
+    for i, j in enumerate(args.args):

[Jim Stichnoth, 2016/01/13 00:36:44]

You're not using "i", right?  I think the typical pattern for this is something
like:

  for _, arg in enumerate(args.args):

[sehr, 2016/01/13 18:15:35]

Done.

+      if re.search('^-?-o(=.+)?', j):

[Jim Stichnoth, 2016/01/13 00:36:44]

I think I would just use a single big regex here, and if it matches, just report
that the arg should be used before --args instead of after, and don't bother
trying to translate between pnacl-sz args and this script's args.

In bash regex, it would be something like:
  {-,--}{o,output,filetype,target,sandbox}{=,}

[sehr, 2016/01/13 18:15:35]

Done.

+        scriptname = os.path.basename(sys.argv[0])

[Jim Stichnoth, 2016/01/13 00:36:44]

Hoist this out of the loop, and remove duplication below.

[sehr, 2016/01/13 18:15:35]

Done.

+        print 'Output file should be set using:'
+        print '    %s ... --output <outfile> ... --args' % scriptname
+        print 'rather than:'
+        print '    %s ... --args -o <outfile>' % scriptname
+        exit(1)
+      elif re.search('^--filetype(=.+)?', j):
+        scriptname = os.path.basename(sys.argv[0])
+        print 'Output file type should be set using:'
+        print '    %s ... --filetype=<type> ... --args' % scriptname
+        print 'rather than:'
+        print '    %s ... --args --filetype=<type>' % scriptname
+        exit(1)
+      elif re.search('^--target(=.+)?', j):

[Jim Stichnoth, 2016/01/13 00:36:44]

Add a similar check for --sandbox

[sehr, 2016/01/13 18:15:35]

Done.

+        scriptname = os.path.basename(sys.argv[0])
+        print 'Target architecture should be set using:'
+        print '    %s ... --target=<target> ... --args' % scriptname
+        print 'rather than:'
+        print '    %s ... --args --target=<target>' % scriptname
+        exit(1)
+    asm_temp = None
+    output_file_name = None
+    keep_output_file = False
+    if args.output:
+      output_file_name = args.output
+      keep_output_file = True
     cmd += args.args
     if args.llvm_source:
       cmd += [llfile]
-    asm_temp = None
     if args.assemble or args.disassemble:
-      # On windows we may need to close the file first before it can be
-      # re-opened by the other tools, so don't do delete-on-close,
-      # and instead manually delete.
-      asm_temp = tempfile.NamedTemporaryFile(delete=False)
-      asm_temp.close()
+      if not output_file_name:
+        # On windows we may need to close the file first before it can be
+        # re-opened by the other tools, so don't do delete-on-close,
+        # and instead manually delete.
+        asm_temp = tempfile.NamedTemporaryFile(delete=False)
+        asm_temp.close()
+        output_file_name = asm_temp.name
     if args.assemble and args.filetype != 'obj':
       cmd += (['|', os.path.join(pnacl_bin_path, 'llvm-mc')] +
               TargetAssemblerFlags(args.target, args.sandbox) +
-              ['-filetype=obj', '-o', asm_temp.name])
-    elif asm_temp:
-      cmd += ['-o', asm_temp.name]
+              ['-filetype=obj', '-o', output_file_name])
+    elif output_file_name:
+      cmd += ['-o', output_file_name]
     if args.disassemble:
       # Show wide instruction encodings, diassemble, and show relocs.
       cmd += (['&&', os.path.join(pnacl_bin_path, 'le32-nacl-objdump')] +
               args.dis_flags +
               ['-w', '-d', '-r'] + TargetDisassemblerFlags(args.target) +
-              [asm_temp.name])
+              [output_file_name])
 
     stdout_result = shellcmd(cmd, echo=args.echo_cmd)
     if not args.echo_cmd:
       sys.stdout.write(stdout_result)
-    if asm_temp:
-      os.remove(asm_temp.name)
+    if asm_temp and not keep_output_file:
+      os.remove(output_file_name)
 
 if __name__ == '__main__':
     main()