Speed up a bit by threading

$ time fix-offsets -p1 ../nongit1.patch > ../nongit2.patch

before:

> real	0m0.472s
> user	0m0.202s
> sys	0m0.262s

after:

> real	0m0.296s
> user	0m0.188s
> sys	0m0.217s

Signed-off-by: Jan Pokorný <jpokorny@redhat.com>

1 files changed, 50 insertions, 16 deletions

diff --git a/fix-offsets b/fix-offsets
index 75b08f9..d9417b5 100755
--- a/fix-offsets
+++ b/fix-offsets
@@ -9,6 +9,7 @@ import re
 import getopt
 from os import environ
 from subprocess import Popen, PIPE
+from threading import Thread
 
 # svn checkout http://python-patch.googlecode.com/svn/trunk/ python-patch
 # pushd python-patch
@@ -23,6 +24,7 @@ re_h = re.compile(
     re.MULTILINE
 )
 
+use_threading = True
 
 def adjustment(ret, stderrdata):
     #print >>sys.stderr, "ret = {0}\nstderrdata = {1}".format(ret, stderrdata)
@@ -35,12 +37,33 @@ def adjustment(ret, stderrdata):
     return ret
 
 
+def hunk_worker(hunk, header, tres, cmd, **kwargs):
+    proc = Popen(cmd, **kwargs)
+    str_hunk = str(hunk)
+    partial_patch = header + '\n' + str_hunk
+    _, stderrdata = proc.communicate(partial_patch)
+    delta = adjustment(proc.wait(), stderrdata)
+    if delta:
+        hunk.startsrc += delta
+        hunk.starttgt += delta
+        str_hunk = str(hunk)
+    tres[hunk] = str_hunk
+
+
 def proceed(opts, args):
-    null = open('/dev/null', 'a')
     if not args:
         args.append('-')
     cmd_args = reduce(lambda a, x: a + list(x), opts, [])
     #print >>sys.stderr, "cmdargs: {0}, opts: {1}".format(cmd_args, opts)
+
+    null = open('/dev/null', 'a')
+    cmd = (lambda *a: a)('/usr/bin/patch', '-o-', *cmd_args)
+    e = dict(LC='C')
+    for name, value in environ.iteritems():
+        if name in ('TMPDIR', 'TMP', 'TEMP'):
+            e[name] = value
+    kwargs = dict(stdin=PIPE, stdout=null, stderr=PIPE, env=e)
+
     for arg in args:
         arg = arg if arg != '-' else '/dev/stdin'
         ps = patch.fromfile(arg)
@@ -51,22 +74,33 @@ def proceed(opts, args):
             header = '{0}--- {1}\n+++ {2}'.format('\n'.join(p.header),
                                                     p.source, p.target)
             print header
+            tres = dict()
+            ts = set()
+            tmax = 10
             for hunk in p.hunks:
-                cmd = (lambda *a: a)('/usr/bin/patch', '-o-', *cmd_args)
-                e = dict(LC='C')
-                for name, value in environ.iteritems():
-                    if name in ('TMPDIR', 'TMP', 'TEMP'):
-                        e[name] = value
-                proc = Popen(cmd, stdin=PIPE, stdout=null, stderr=PIPE, env=e)
-                str_hunk = str(hunk)
-                partial_patch = header + '\n' + str_hunk
-                _, stderrdata = proc.communicate(partial_patch)
-                delta = adjustment(proc.wait(), stderrdata)
-                if delta:
-                    hunk.startsrc += delta
-                    hunk.starttgt += delta
-                    str_hunk = str(hunk)
-                print str_hunk,
+                if not use_threading:
+                    hunk_worker(hunk, header, tres, cmd, **kwargs)
+                    continue
+                if len(ts) >= tmax:
+                    while True:
+                        for t in ts:
+                            t.join(0.001)
+                            if not t.is_alive():
+                                ts.remove(t)
+                                break
+                        else:
+                            continue
+                        break
+                t = Thread(target=hunk_worker, args=(hunk, header, tres, cmd),
+                                                     kwargs=kwargs)
+                t.start()
+                ts.add(t)
+            if use_threading:
+                for t in ts:
+                    t.join()
+            for hunk in p.hunks:
+                print tres[hunk],
+
     null.close()